RightChain Nodes User Guide

RightChain™ Nodes Supply Chain Network Optimization powered by RightChain.ai 1. Introduction 2. Login Instructions 3. Navigation 4. Working with Data 5. Modeling 6. Plots, Charts, and Visualizations 7. Methodology 8. Messages and Icons

RightChain™ Nodes User Guide | © RightChain Inc.

1. Introduction RightChain Nodes is an AI-based supply chain network optimization system. It is offered on the RightChain.ai platform as subscription-based software-as-a-service. The underlying algorithms, methodology, and visualizations are the result of decades of development by RightChain's optimization technology team including Dr. Ed Frazelle, founder of the Supply Chain Logistics Institute at Georgia Tech; Dr. Travis Smith, a literal rocket scientist holding a doctorate of aerospace engineering from Georgia Tech; and Dr. Andrew Frazelle, a PhD in decision sciences from Duke University and a member of the faculty of the prestigious Jindal School of Management at the University of Texas Dallas. RightChain Nodes has been employed in crafting the supply chain networks and strategies of many of the world's most successful supply chains including those at adidas, BP, Carrier, Caterpillar, Coca-Cola, Dell, Disney, Fisher Scientific, Nutrisystem, Pratt & Whitney, United Technologies, and UPS among many others. The RightChain Nodes Help System includes sections on FAQs, Logins, Application Navigation, Working with Data, Modeling, Visuals and Plots, Methodology, and Use Cases. The help system in continually being updated based on input from users and inspirations from the development team.

RightChain™ Nodes User Guide | © RightChain Inc.

Page 2 of 57

2. Login Instructions Please follow the instructions below to login to RightChain Nodes.

RightChain.ai Login Screen. All users receive an invitation email and password.

RightChain Nodes

Page 3 of 57

3. Navigation 3.1 Page Types 3.2 Page Navigation 3.3 Sidebar Interactions

RightChain Nodes

Page 4 of 57

3.1 Page Types RightChain.ai hosts and operates two distinctly different page types. PAGE TYPE 1 (ALL CAPS) hosts and executes Analytics, Optimizations, and Visualizations. There is only one Page Type 1 in a RightChain.ai application. PAGE TYPE 1 is selectable in the upper left hand corner of all RightChain.ai applications. Page Type 2 (Upper and Lower Case) hosts and executes Data Inputs. There may be one or more Page Type 2's in any given application, depending upon the data requirements for the application. Page Type 2's are selectable from the upper left hand side of any page within an application. Example PAGE TYPE 1 and Page Type 2 are depicted below.

An example of RightChain.ai Page Type 1 hosting and executing analytics, optimizations, and visualizations.

An example of RightChain.ai Page Type 2 hosting and executing data inputs.

RightChain Nodes

Page 5 of 57

3.2 Page Navigation Each of the two types of RightChain.ai pages are composed of functional sectors. The functional sectors for PAGE TYPE 1 are illustrated below. PAGE TYPE 1's are composed of ten sectors.

SECTOR ID

SECTOR NAME

NOTES

Home to the platform name, "RightChain.ai". Click on "RightChain.ai" to access the RightChain.ai home page. Sector A is also home to User Credentials and "Sign Out". IN ALL CAPS, THE APPLICATION NAME IS DISPLAYED. Click on the application name to access the application's main analytical functions. Each RightChain.ai application works with two page types; an APPLICATION PAGE (PAGE TYPE 1), and Data Input pages (Page Type 2). Select "Input: abc" to enter data The Application Navigation sector includes the name of the RightChain.ai suite and a drop down menu allowing users to switch between modules within the active suite. The File Management sector is home to access the saved file directories, file loading and uploading, and file saving. The Optimization Scenario Specification sector allows user to specify optimization scenarios via selecting input filters, modifying optimizaiton parameters, and choose/modify solution plots. The Application Visualization Tabs sector houses the tab names created by users with each tab representing a unique visualization. The Solution Visualization sector houses the optimization visualization and/or plots related to the optimization solution and/or its data. The Solution Evaluation and Statistics sector houses any and all statistics and performance measures related to the optimization solution and/or associated plots. The Solution Notes and Assumptions allows users to record and save notes and/or assumptions related to the optimization solution and/or related plots.

Sector A

Platform Navigation Ribbon

Sector B

APPLICATION NAME

Sector C

Data Inputs

Sector D

Application Navigation

Sector E

File Management

Sector F

Optimization Scenario Specification

Sector G

Application Visualization Tabs

Sector H

Solution Visualization

Sector I

Solution Evaluation and Statistics

Sector J

Solution Notes and Assumptions

Table of Page Type 1 Sectors

RightChain Nodes

Page 6 of 57

Page Type 1 Sectors

RightChain Nodes

Page 7 of 57

3.3 Sidebar Interactions Content in the left sidebar (labeled: Sidebar Interactions) is available to filter data, configure parameters, and customize the data visualizations that are part of each Module Subtab. The interactions will change depending on the subtab that you are currently viewing. This is frequently true for the “Plot Controls” which provide numerous options for configuring visualizations. Clicking on the interaction headings, in bold (like “Input Filters”), expands that interaction group to reveal what modifications can be made. The screenshot below shows the “Input Filters” interactions expanded which can be used include or excluded specific data filters and apply them to all of the module subtabs and plots on the right side of the screen.

RightChain Nodes

Page 8 of 57

4. Working with Data 4.1 Uploading New Data 4.2 Saving Results

4.3 Loading Previously Saved Data 4.4 Data Templates for Data Uploads 4.5 Data Types

RightChain Nodes

Page 9 of 57

4.1 Uploading New Data New data is uploaded on the “Input: Data Set” screen. From this screen, you can also download data templates, assign columns to required variables for the module, and change the data type or RightChain variable associated with a specific column in the dataset. To upload a new data set from your computer, click on the “Browse…” button – under the Upload Order Data” heading – and locate a file that is one of the approved file types. Currently, the accepted file types include Excel files (.XLS, .XLSX) and CSV files (.CSV).

After loading, the data set will display in the main panel with its automatically assigned data types and RightChain Variables. The next section will discuss more about how to use the dropdown boxes to assign data types and RightChain Variables.

RightChain Nodes

Page 10 of 57

RightChain Nodes

Page 11 of 57

4.2 Saving Results The process for saving data set and module data is very similar to the Loading Saved Data instructions. New data sets, configured plots, or analyses run can be saved individually or together by clicking the file saving icons in the top right menu area of the application. The “Save As” button displays saving options to select the folder the file should be saved in, the file name, and checkboxes to decide if the data set, analysis data, and plots will all be saved or just some of them (screenshot below). Saved files can also be designated as “Read Only” if users want the file's creator to be able to save over it.

The “Save All” button, on the far right of the menu bar, saves all parts of the application (input data, results, and plots). After saving has completed, the saved data will be available to load for this module and others.

RightChain Nodes

Page 12 of 57

4.3 Loading Previously Saved Data Saved datasets from previous sessions can be accessed via the file loading icon in the top right menu area of the application. From there, files saved in your directory (or the “Demo Files” directory) can be selected and loaded by clicking the “Load Selected” button.

When a file is selected, check boxes appear at the bottom of the panel which give users the option to load just one part of or all of the saved file. When you have determined what you want to load, click the “Load Selected” button, and a confirmation screen will appear. Next, the loaded data will be displayed in the module subtabs, and the “Input: Data Set” will display the data set and allow you to assign columns to required variables and modify variable types if needed (instructions found in Setting Data Types and RightChain Variables ).

RightChain Nodes

Page 13 of 57

4.4 Data Templates for Data Uploads When uploading new data, data template files provide a sample of the column names, data types, and structure of accepted input files for ease of uploading new or refreshed data. Some modules have template files available with column names pre-populated available on the left sidebar of the “Input: Data Set” page via the Download button (shown below).

RightChain Nodes

Page 14 of 57

4.5 Data Types After a data set is loaded or uploaded, a table displaying the data will also appear with selection boxes for assigning data types and RightChain variables to each column. A list of required variables that are specific to each module is shown in the “Required” group at the top. All of the required variables are necessary to take full advantage of available features and analyses within an application. The application automatically identifies variables in the data set that could be these required variables as well as data types for the columns. If the automatic assignment does not match the data set exactly, user can use the dropdown boxes to make the appropriate changes.

Since the RightChain Variable boxes contain many possible options, they are easily searchable by typing into the selection box. The required inputs can either be set with the dropdown below their Original Column name in the table or using the list on the left sidebar. The image below shows how the RightChain Variable for the “SKU” column can be set to one of the Required Inputs

RightChain Nodes

Page 15 of 57

of “Eaches”, “SKU”, “Customer ID”, or “Warehouse Order Number” from its column dropdown or in the four “Required Inputs:” dropdown boxes in the left sidebar.

When finished uploading the dataset and assigning variables, clicking on the application name in the top left corner will return the AI to the main page with all of its corresponding subtabs available for working with the data.

RightChain Nodes

Page 16 of 57

5. Modeling 5.1 Modeling New Locations 5.2 Transportation Cost Modeling

5.3 Geographic Identifiers 5.4 User Defined Locations 5.5 Fixed Locations 5.6 Network Consolidation 5.7 Network KPIs 5.8 Incremental Inventory 5.9 Distance Modeling 5.10 Inbound and Outbound Transportation

RightChain Nodes

Page 17 of 57

5.1 How do I add a NEW WAREHOUSE to the existing network? To define the existing network, users upload the existing warehouses into “User Defined Locations”. Once completed, users can select from among the existing warehouses from the “User Defined Locations” selection under the “Parameters” section. At that point, users may set the number of DCs to solve for to the existing number plus one. Please make sure to turn on the “Use Pre-Defined Locations” switch when working with this scenario.

RightChain Nodes

Page 18 of 57

5.2 Transportation Cost Modeling RightChain Nodes uses a cost minimization approach to find the optimal facility locations based on the “Location Contribution Factor” and the distance to each location. Transportation costs may be computed using a "universal rate" or a "rate table". Additional rate formats may be made available upon request. Universal Rates The universal rate assumes that shipping activity can be multiplied by a constant number to compute the transportation cost. For example, shipping 1lb over 100 miles at a rate of $0.20 per lb-mile would cost $20, while shipping 5lbs over 5 miles would cost $4. Universal rates depend upon many factors including the transportation mode (full truckload, less-than truckload, parcel, air, ocean, rail, etc.), the unit of measure for shipping activity (weight, cube, distance, units, etc.), and prevailing/negotiated transportation rates. RightChain has employed its universal rate approach very successfully in nearly every major industry, geography, and mode of transportation. Lane Rates Lane Rates assume that each lane either has a fixed cost or a similar structure to the universal rate calculations. This transportation costing approach assumes that the “Transportation Rate Table” has been uploaded and that all possible lanes have been enumerated, and rated.

Example transportation cost computation using a universal rate.

RightChain Nodes

Page 19 of 57

5.3 What are considered valid GEOGRAPHIC IDENTIFIERS? RightChain Nodes provides select support for geocoding within the AI. RightChain Nodes supports geocoding for the “Postal Code”, “State/Region/Province”, and “Country” levels. RightChain.ai supports geocoding for a select set of countries. Supported countries are listed under the “Default Country” selection under the “Data Configuration” section within the AI. This list may be updated at any point, and additional data may be added upon request. (RightChain recommends that users provide the “Latitude” and “Longitude” values in their data sets for more precise results.) Additionally, as different countries have different levels of accuracy with their postal codes, (Canadian postal codes specify the side of the street rather than a region), RightChain may truncate postal codes and use only the first 3 characters. The postal codes by country supported by RightChain are available within the “Presets and Postal Code Locations” selection under the “Parameters” section within the AI.

RightChain Nodes

Page 20 of 57

5.4 What are “User-Defined Locations”, and what are they used for? “User-Defined Locations” are locations of interest defined by the user that the AI can use as a fixed location(s) when optimizing. Simply, "user-defined locations" allow users to define a point in the solution for consideration as a potential or required optimal location. Once a user has uploaded the user-defined locations, additional selections are made available under the “Data Configuration” and “Parameters” sections of “Optimization Controls” within the AI. To model your existing network, you can upload your existing DCs in the “User-Defined Locations” data sheet, and follow the instructions concerning fixing locations within the solution.

RightChain Nodes

Page 21 of 57

5.5 How do I FIX LOCATIONS within a solution? The AI allows users to pre-define or fix in-place a location(s) within the solution. In order to do so, users must switch on “Use Pre-Defined Locations” within the “Parameters” section. Users can then choose their locations from either the “Presets and Postal Code Locations” selection or the “User Defined Locations” selection. If there are fewer pre-defined locations selected than the number of DCs solved for, then the remaining locations will be solved. If there are more pre-defined locations selected than the number of DCs solved for, then the AI will only use the first X pre-defined locations. This allows users to develop and execute different scenarios within the AI.

Users may "pre-define" or "fix in place" specific locations for the AI to consider in developing optimal solution designs.

RightChain Nodes

Page 22 of 57

5.6 How do I CONSOLIDATE warehouses within the network? To define the existing network, users upload the existing warehouses into the “User Defined Locations” data input. Once complete, users can select the existing warehouses from the “User Defined Locations” selection under the “Parameters” section minus the warehouses under consideration for consolidation. At that point, users set the number of the DCs to solve for to the desired number. Please remember to turn on the “Use Pre-Defined Locations” switch when working with this scenario.

RightChain Nodes

Page 23 of 57

5.7 What are the KEY COST, SERVICE, AND PERFORMANCE FACTORS monitored for supply chain network performance in RightChain Nodes?

FUTURE DCs 1. Future DCs . The target number of distribution centers or warehouses in the optimal solution.

SERVICE TIMES 1. Half Day. The percent of the "activity" delivered within a half-day. The unit of measure for "activity" is selected by the user, and typically is a unit of measure like weight, cube, units, or lines. 2. Same Day. The percent of the "activity" delivered on the same day. The unit of measure for "activity" is selected by the user, and typically is a unit of measure like weight, cube, units, or lines. 3. 1 Day. The percent of the "activity" delivered within 24 hours. The unit of measure for "activity" is selected by the user, and typically is a unit of measure like weight, cube, units, or lines. 3. 2 Days. The percent of the "activity" delivered within two days. The unit of measure for "activity" is selected by the user, and typically is a unit of measure like weight, cube, units, or lines. COSTS 1. Transport Cost. The total cost of transportation based upon the type of rate selected (universal or table) and it's corresponding unit of measure (e.g. miles, pounds, cube, pound miles, etc.). 2. I nventory Carrying Cost. The cost of "carrying" the network's inventory investment; computed as the product of the inventory carrying rate specified in the scenario and the network's inventory investment. 3. Total Logistics Cost. The sum of the scenario's transportation cost and inventory carrying cost. 4. Footprint. The estimated square footage occupied by facilities in the network, computed as the product of the network's inventory investment and the storage density of the facilities. 5. Inventory Investment. The estimated inventory in the network based upon the number of network facilities and the portion of inventory allocated to safety stock, lot size, and pipeline inventory. 6. % Incremental Inventory . The estimated increase or decrease in the network's inventory investment compared to the baseline inventory investment.

RightChain Nodes

Page 24 of 57

Solution KPIs include service times, transportation costs, inventory carrying costs, inventory levels, and network footprint.

RightChain Nodes

Page 25 of 57

5.8 Incremental Inventory The incremental inventory investment is an estimate of the change in inventory investment required throughout the network as a function of the number of DCs, warehouses, or stocking locations. The total inventory investment is based on the “Inventory Parameters” and the total number of distribution centers. As the number of distribution centers increases, the required amount of safety stock throughout the network increases. RightChain Nodes automatically takes that phenomenon into account when computing the inventory investment required for a given network scenario. In the example, there are three facilities in the baseline network, with a total of $100,000,000 inventory investment. In increasing the number of network facilities from three to four, RightChain Nodes estimates that there will be an additional 7.74% inventory, or $7,740,000.

Baseline Scenario with Three DCs

RightChain Nodes

Page 26 of 57

Four DC Scenario with Incremental Inventory

RightChain Nodes

Page 27 of 57

5.9 Distance Modeling RightChain.ai models the distance between locations using the Haversine distance computation modified by an index based upon modes of transportation taking into account road, rail, ocean, and air routing. The Haversine distance, often referred to as the Haversine formula or Haversine formula distance, is a method used to calculate the distance between two points on the surface of a sphere, such as the Earth. It is particularly useful for computing distances between locations specified by their latitude and longitude coordinates. The formula is named after its inventor, R. W. Sinnott, who published it in 1984, although the underlying principles have been known for centuries. The Haversine formula is based on spherical trigonometry and is particularly accurate for short to moderate distances on the Earth's surface. It calculates the great-circle distance, which is the shortest distance between two points on a sphere's surface along the surface of the sphere (like the arc of a circle). This makes it suitable for measuring distances over the curved surface of the Earth, as opposed to linear distance calculations that assume a flat surface. The Haversine formula is commonly used in geographic applications, such as mapping, geolocation, and navigation systems, to calculate distances between GPS coordinates or other latitude and longitude pairs on the Earth's surface. It provides reasonably accurate results for most practical purposes, but it should be noted that the Earth's shape is not a perfect sphere, so more complex geodesic distance formulas may be employed for high-precision calculations over longer distances or in specific geographic regions.

RightChain Nodes

Page 28 of 57

5.10 Inbound vs Outbound RightChain Nodes treats Inbound and Outbound shipments separately. The outbound shipments are mapped to the “Ship To Location Basis” while the Inbound shipments are mapped to the “Ship From Location Basis”. For the application to recognize the difference between the two, a column with the values “Inbound” and “Outbound” is required within the data set.

Copyright: RightChain Incorporated | All Rights Reserved

RightChain Nodes

Page 29 of 57

6. Plots, Charts, and Visualization 6.1 Bar Charts 6.2 Pareto Charts 6.3 Histograms

6.4 XY Charts 6.5 Partitions 6.6 Outliers 6.7 Predictions 6.8 Networks 6.9 Correlations 6.10 Maps 6.11 Data Point Markers

RightChain Nodes

Page 30 of 57

6.1 Bar Charts Bar Charts are one type of many types of visualizations available to users in RightChain.ai. Bar charts display the contribution of observations to the total, with observations ranked at the users discretion. In this example, LAT-LON coordinates are displayed with cubic volume shipped to each coordinate represented in the height of the bar.

Bar Chart Example

RightChain Nodes

Page 31 of 57

Bar Chart Description

A bar chart, also known as a bar graph, is a graphical representation of data that uses rectangular bars to display and compare the values of different categories or data sets. Bar charts are widely used for visualizing categorical or discrete data and are effective for showing relative comparisons between items. Key characteristics of a bar chart include: 1. Categories or Data Sets: The horizontal axis of the bar chart represents categories or data sets being compared. These categories can be anything from time periods, products, regions, or any other discrete labels. 2. Vertical Bars: For each category on the horizontal axis, a vertical bar is drawn. The height or length of each bar represents the value or quantity associated with that category. The taller the bar, the greater the value. 3. Values or Counts: The vertical axis, often referred to as the y-axis, represents the values or counts associated with the categories. The scale of the y-axis depends on the range of values in the data. 4. Spacing: Typically, there is a small gap between adjacent bars in a bar chart, which visually separates them. However, you can create clustered bar charts where bars of the same category group together without gaps, or stacked bar charts where bars are stacked on top of each other to show the composition of a whole. Bar charts can come in different variations, depending on the nature of the data and the specific information you want to convey. Some common types of bar charts include:

RightChain Nodes

Page 32 of 57

1. Vertical Bar Chart: The bars are oriented vertically, with categories along the horizontal axis. This is the most common type of bar chart. 2. Horizontal Bar Chart: The bars are oriented horizontally, with categories along the vertical axis. Horizontal bar charts are useful when you have long category labels or when you want to emphasize the differences in values more easily. 3. Grouped Bar Chart: Also known as clustered bar charts, these display multiple bars for each category side by side, allowing for easy comparison between different data sets within each category. 4. Stacked Bar Chart: In a stacked bar chart, bars for each category are stacked on top of each other, representing the total value for each category. This type of chart is useful for illustrating the composition of a whole and the contribution of each component. Bar charts are commonly used in various fields and scenarios, including business, economics, marketing, education, and research, to visualize data and make comparisons between different categories or groups. They are particularly effective for displaying data that can be divided into distinct categories and for showing trends or differences among those categories.

RightChain Nodes

Page 33 of 57

6.2 Pareto Charts PARETOS are one of many types of visualizations available to users in RightChain.ai. Pareto charts display the cumulative contribution of increasing numbers of observations to the total, with observations ranked from highest to lowest value. In this example, LAT-LON coordinates are ranked from high to low based upon the cubic volume shipped to each coordinate.

Example Pareto Plot

RightChain Nodes

Page 34 of 57

Pareto

Pareto Plot Controls

A Pareto chart, also known as a Pareto diagram, is a graphical tool used for analyzing and prioritizing the most significant factors within a dataset. It is named after Vilfredo Pareto, an Italian economist and sociologist who observed that a significant portion of wealth is typically owned by a small percentage of the population. The Pareto chart is commonly used in quality management and process improvement to identify and address the most critical issues or causes of problems. Here are the key components and characteristics of a Pareto chart: 1. Vertical Bar Chart: A Pareto chart is typically presented as a vertical bar chart. The vertical axis represents the frequency or count of occurrences, while the horizontal axis displays the categories or factors being analyzed. 2. Categories or Factors: The categories or factors are listed on the horizontal axis. These can represent different aspects, such as types of defects, causes of issues, product features, or customer complaints. The categories are usually arranged in descending order based on their frequency or impact. 3. Bars: For each category or factor, a vertical bar is drawn, with the height of the bar representing the frequency or count of occurrences associated with that category. The categories with the highest bars are the most significant or impactful. 4. Cumulative Percentage: A Pareto chart often includes a line graph (a cumulative percentage line) that shows the cumulative percentage of the total occurrences as you move from left to right along the categories. This line helps identify the point at which a certain percentage of the total is reached, highlighting the most critical categories. The primary purpose of a Pareto chart is to identify the "vital few" from the "trivial many." In other words, it helps you focus your efforts and resources on addressing the most important

RightChain Nodes

Page 35 of 57

issues or factors that have the greatest impact on a problem or goal. By doing so, you can achieve more significant improvements or solutions efficiently. Pareto charts are widely used in various fields, including: • Quality Control: In quality management, Pareto charts are used to identify the most common defects or problems in a production process so that corrective actions can be prioritized. • Project Management: In project management, Pareto charts can be used to identify the most common sources of delays or issues in a project. • Customer Complaints: In customer service and satisfaction analysis, Pareto charts can help identify the most frequent reasons for customer complaints, allowing organizations to address them proactively. • Product Development: In product design and development, Pareto charts can be used to prioritize features or issues that need attention based on customer feedback or testing results. Overall, Pareto charts are a valuable tool for making informed decisions, setting priorities, and focusing resources on the areas that will have the most significant impact on achieving desired outcomes.

RightChain Nodes

Page 36 of 57

6.3 Histograms HISTOGRAMS are one type of many types of visualizations available to users in RightChain.ai. Histograms display the portion of observations within a given range of values. In this example, the portion of LAT-LON coordinates within ranges of buckets of cubic volume shipped to each coordinate are represented in the height of the bar.

Example Histogram Plot

RightChain Nodes

Page 37 of 57

Histogram Plot Controls

A histogram is a graphical representation of the distribution of a dataset. It is a way to visualize the underlying frequency or probability distribution of a set of continuous or discrete data. Histograms are commonly used in statistics and data analysis to understand the shape, central tendency, and spread of a dataset. Here are the key components and characteristics of a histogram: 1. Bins or Intervals: A histogram divides the range of data into a set of non-overlapping intervals, also known as "bins" or "buckets." These bins represent different ranges of values within the dataset. The choice of bin width can affect the appearance of the histogram and the insights it provides. 2. Frequency or Count: For each bin, the histogram displays the frequency or count of data points that fall within that bin's range. In other words, it shows how many data points belong to each interval. 3. Bars: The histogram is constructed by drawing vertical bars above each bin, with the height of each bar proportional to the frequency or count of data points in that bin. Taller bars indicate a higher concentration of data points in that interval. 4. X-Axis: The x-axis of the histogram represents the data values or the ranges of values. It is divided into the bins, and each bin is labeled or marked along the x-axis. 5. Y-Axis: The y-axis of the histogram represents the frequency or count of data points in each bin. The scale of the y-axis depends on the range of data and the number of data points.

RightChain Nodes

Page 38 of 57

Histograms are particularly useful for the following purposes: • Visualizing Data Distribution: Histograms provide a visual summary of how data is distributed. They can reveal patterns such as skewness (whether the data is positively or negatively skewed) and modality (whether it has one or multiple peaks). • Identifying Outliers: Outliers or extreme values can often be identified in a histogram as data points that fall far from the main concentration of data. • Selecting Appropriate Data Analysis Techniques: Understanding the distribution of data can help in selecting appropriate statistical analysis techniques, such as choosing between parametric and non-parametric methods. • Data Preprocessing: In data preprocessing, histograms are used to understand the distribution of variables and make decisions about data transformation, normalization, or outlier removal. Histograms are a fundamental tool in exploratory data analysis and are frequently used in fields such as statistics, data science, and quality control to gain insights into the characteristics of datasets. They provide a visual representation of data that can be easily interpreted and analyzed.

RightChain Nodes

Page 39 of 57

6.4 XY Plots XY Plots, sometimes referred to as "Scatter Plots", are two-dimensional depictions of observational statistics; one statistic on the X axis, and the other statistic on the Y axis.

Example XY Plot

XY Plot Controls

RightChain Nodes

Page 40 of 57

XY Charts An XY chart, also known as a scatter plot or scatter chart, is a type of data visualization used in statistics and data analysis. It displays individual data points as dots or markers on a two dimensional coordinate system, with one variable plotted on the horizontal (x) axis and another variable plotted on the vertical (y) axis. Each data point's position on the chart is determined by its x and y values. Here are some key characteristics and uses of XY charts: 1. Individual Data Points: Unlike other types of charts like line charts or bar charts, XY charts do not connect data points with lines or bars. Instead, each data point is represented independently as a marker on the chart. 2. Continuous or Numerical Data: XY charts are typically used to visualize the relationship between two continuous or numerical variables. For example, you might use an XY chart to show the relationship between temperature (x-axis) and ice cream sales (y-axis) over a period of time. 3. Scatter Patterns: The pattern or distribution of data points on an XY chart can reveal insights about the relationship between the two variables. Common patterns include linear relationships, clusters, curves, and outliers. 4. Correlation Analysis: XY charts are frequently used in correlation analysis to determine if there is a relationship between the two variables. Positive correlation, negative correlation, or no correlation can be observed by examining the direction and strength of the relationship between the data points. 5. Data Exploration: XY charts are useful for exploring and visualizing data, identifying trends, spotting anomalies, and making data-driven decisions. 6. Data Labels: Data points on an XY chart can be labeled with additional information to make the chart more informative. 7. Trendlines: In some cases, you can add trendlines (such as linear regression lines) to an XY chart to visualize the overall trend or pattern in the data. XY charts are commonly used in fields such as science, engineering, finance, social sciences, and data analysis to analyze and present data relationships. They provide a versatile and effective way to visualize and interpret the relationships between two variables, making them a valuable tool in data exploration and hypothesis testing.

RightChain Nodes

Page 41 of 57

6.5 Partitions PARTITIONS are another means of displaying key and related data points in RightChain.ai. In this case, the user has selected three DCs for analysis and insight.

Example Partitions Plot

Partitions Plot Controls

RightChain Nodes

Page 42 of 57

6.6 Outliers “Outliers” are one of many types of visualizations available to users in RightChain.ai. Outliers display the distribution of data points relative to range of “normalcy”.

Example Outliers Plot

Outliers Plot Controls

RightChain Nodes

Page 43 of 57

6.7 Predictions “PREDICTIONS” allow users to create traditional and AI-based activity forecasts for any data for which dates are provided.

Example Predictions Plot

Predictions Plot Controls

RightChain Nodes

Page 44 of 57

6.8 Networks “Networks” are one of many types of visualizations available to users in RightChain.ai. Networks display the activity between connected points in a data set.

Example Network Plot

Networks Plot Controls

RightChain Nodes

Page 45 of 57

6.9 Correlations “Correlations” are one of many types of visualizations available to users in RightChain.ai. Correlations display the strength of association between pairs of factors or objects. In the example the strength of association between pairs of factors (weight, units, sales, lines, and cubic volume) are displayed related to a distribution center.

Example Correlations Plot

Correlations Plot Controls

RightChain Nodes

Page 46 of 57

6.10 Maps MAPS are one type of visualization creatable via RightChain.ai optimization and insights. MAPS require geographic data in the core input data set. Those types of geographic data include latitudes, longitudes, postal codes, cities, states, and countries.

Mapping Example

Maps Plot Controls

RightChain Nodes

Page 47 of 57

6.11 Data Point Markers Data points in plots may take one or more of the following attributes. Data point colors may be based on any of the data columns in the input data set. Data point colors are automatically assigned from a color pallete selected by the user. In addition, data points may be sized or shaped based upon any of the data columns in the input data set.

Data point marker attributes.

RightChain Nodes

Page 48 of 57

7. Methodology 7.1 Clustering 7.2 Optimization

RightChain Nodes

Page 49 of 57

7.1 Clustering RightChain.ai makes extensive use of clustering in crafting visualizations and optimization solutions. Clustering is a machine learning and data analysis technique used to group similar data points or objects into clusters or groups based on their intrinsic similarities. The goal of clustering is to discover meaningful patterns or structures in the data, which can be useful for various purposes, including data exploration, pattern recognition, and data segmentation.

Example RightChain.ai clustering classification of customers as A (blue), B (green), C (yellow), and D (red) based on their net income per order, net incomer per order percent, and revenue per order.

Below are some key points about clustering:

RightChain Nodes

Page 50 of 57

1. Unsupervised Learning: Clustering is typically an unsupervised learning technique, which means that it does not rely on labeled data with predefined categories or classes. Instead, it identifies patterns in the data without prior knowledge of what those patterns might be. 2. Similarity Measure: Clustering algorithms use a similarity measure or distance metric to assess how similar or dissimilar data points are to each other. Common distance measures include Euclidean distance, cosine similarity, and Jaccard similarity, depending on the type of data and the problem at hand. 3. Clustering Algorithms: There are various clustering algorithms available, each with its own approach and characteristics. Some popular clustering algorithms include K-Means, Hierarchical Clustering, DBSCAN (Density-Based Spatial Clustering of Applications with Noise), and Gaussian Mixture Models (GMM). The choice of algorithm depends on the nature of the data and the goals of the analysis. 4. Number of Clusters: In many clustering algorithms, you need to specify the number of clusters you want to create. This can be a challenging task, as it depends on your understanding of the data and the problem. Some algorithms, like K-Means, require you to specify the number of clusters in advance, while others, like DBSCAN, can automatically determine the number of clusters. 5. Cluster Interpretation: After clustering, it is essential to interpret the results and understand the characteristics of each cluster. This may involve examining the central tendencies, outliers, or common attributes within each cluster to derive meaningful insights. 6. Applications: Clustering has a wide range of applications across various domains. For example, it can be used in customer segmentation for marketing, document clustering in natural language processing, anomaly detection in cybersecurity, and image segmentation in computer vision. 7. Evaluation: Evaluating the quality of clustering results can be challenging, as it is often subjective and depends on the specific goals of the analysis. Common evaluation metrics include silhouette score, Davies-Bouldin index, and the visual inspection of cluster quality. Clustering is a valuable tool for data exploration and pattern recognition when you have unstructured or unlabeled data. It can help uncover hidden structures within the data and enable further analysis or decision-making based on the identified clusters. However, the effectiveness of clustering depends on the choice of algorithm, parameter settings, and domain knowledge to interpret the results effectively.

RightChain Nodes

Page 51 of 57

7.2 Optimization Optimization refers to the process of finding the best solution or outcome from a set of possible choices or variables, often with the goal of maximizing or minimizing a specific objective. It is a fundamental concept used in various fields, including mathematics, engineering, economics, computer science, and more. The primary aim of optimization is to improve a system, design, process, or decision by making it as efficient or effective as possible under given constraints. Here are some key aspects of optimization: 1. Objective Function: In optimization problems, there is typically an objective function that needs to be either maximized or minimized. This function represents what you want to optimize, whether it's profit, cost, efficiency, performance, or any other measurable quantity. 2. Decision Variables: Optimization involves manipulating a set of decision variables or parameters that influence the objective function. These variables are adjusted to find the best solution. 3. Constraints: Constraints are limitations or restrictions that the solution must satisfy. These constraints can be physical, economic, or other practical limitations that define the feasible solution space. 4. Optimization Algorithms: Various algorithms and techniques are used to search for the optimal solution. The choice of algorithm depends on the nature of the problem, such as linear programming, gradient descent, genetic algorithms, and many others. 5. Local vs. Global Optimization: In some cases, optimization may find a local optimum, which is the best solution within a limited region of the solution space. Finding the global optimum, which is the absolute best solution across the entire space, can be more challenging. 6. Continuous vs. Discrete Optimization: Optimization problems can be classified as continuous or discrete. In continuous optimization, decision variables can take any real value within a given range, while in discrete optimization, variables are limited to specific discrete values. 7. Applications: Optimization is used in various real-world applications, such as in supply chain management, finance, engineering design, machine learning, and logistics, among others. 8. Trade-offs: Optimization often involves trade-offs between conflicting objectives. For example, optimizing cost may lead to a decrease in quality, and vice versa. Balancing these trade-offs is an essential aspect of the optimization process. Overall, optimization is a powerful tool for improving processes, making informed decisions, and achieving the best possible outcomes in a wide range of fields. It involves a combination of mathematical modeling, algorithmic techniques, and problem-solving skills.

RightChain Nodes

Page 52 of 57

8. User Messages and Symbols 8.1 Missing Data Message 8.2 Files Saved Message 8.3 Icons and Symbols

RightChain Nodes

Page 53 of 57

8.1 Missing Data The "Missing Data Set" message may appear when launching an application for which no data has been loaded. Click "OK". The message will disappear. Then load a data set or upload a new data set.

Missing Data Set Message

RightChain Nodes

Page 54 of 57

8.2 Files Saved Message Once saving is complete, RightChain.ai responds with a message verifying the save. Click "Ok" to continue.

RightChain Nodes

Page 55 of 57

8.3 Icons and Symbols RightChain.ai employs a variety of icons and symbols to assist users in navigating the application and in customizing their solutions and related visualizations.

SYMBOL

DESCRIPTION

NOTES

Allows users to modify the appearance of bar charts, histograms, and Pareto charts. Allows users to view data as a chart as opposed to a table.

Bar Type Chart Appearance

Chart

Chart Configuration

Allows users to modify chart types and names.

Chart Deletion

Allows users to delete existing charts, plots, or visulations.

Allows users to define clustering criteria as well as the number of clusters. Clustering is typically used to stratify data points such as Very Heavy, Heavy, Medium, and Light days of the week; or Diamond, Gold, Silver, and Bronze customers. Allows users to configure the display of data points on a plot based upon colors, size, and shape. Allows users to display text with data points on plots including labels and hover text.

Clustering

Data Point Display

Data Point Text

Download Data

Allow users to download solutions or input data.

File Folder

Opens a user's RightChain.ai file folder library.

Allows users to select the rendering speed for graphics. Some graphical detail and resolution may be sacrificed at higher speeds. Allows users to select a previously saved data set from a dropdown directory. Allows users to add a plot, chart, or visualization. New plots, charts, and visualizations may be named on tabs for later viewing. Allows user to "quick save" their current work session and related data. Allows users to save their work and related data with the file name of their choice. This function is often used to name optimization scenarios. Allows users to configure and display statistics in their visualizations. Available statistics include Minimum, 1st Quartile, Median, Mean, 3rd Quartile, Maximum, Correlation, and Kurtosis.

Graphical Performance

Load Files

Plus Sign

"Save"

"Save As"

Statistics

Table

Allows users to view solutions in tabular form.

RightChain™ Nodes User Guide | © RightChain Inc.

Page 56 of 57

Click to reveal additional information or help concerning a particular function, chart, value, or computation.

Tooltip

RightChain Nodes

Page 57 of 57