Presentation is loading. Please wait.

Presentation is loading. Please wait.

Using Veera with R and Shiny to Build Complex Visualizations

Similar presentations


Presentation on theme: "Using Veera with R and Shiny to Build Complex Visualizations"— Presentation transcript:

1 Using Veera with R and Shiny to Build Complex Visualizations
Bharathwaj Vijayakumar, Senior Business Intelligence Analyst, Rowan University #RIConf2019

2 Overview of the presentation
Using Veera with R and Shiny to Build Complex Visualizations Documentation and Veera Workflow

3 Objectives Visualize the number of number of students migrating from one major to another Visualize the sequence of courses that students take in a specific program Visualize the effect of grades in one course, on the courses that follow that

4 Tools and Technology Used
SQL Veera R Shiny R and Shiny was used because Tableau did not directly support the chart types needed for this type of data

5 Visualizations Used and Why?
Sankey Chart Advantages Look at magnitude of a flow from point A to point B Very easy to interpret and visually pleasing Can be created as multilevel, A to B to C Disadvantages Not Excel or table! (Can be an advantage too!) Sometimes flows can looks similar Scale can be hard to interpret sometimes Needs dataset to be in a specific way Cannot Handle cyclic flows (A to B to A)

6 Visualizations Used and Why?
Chord Diagram Advantages Look at magnitude of a flow from point A to point B Visually pleasing Can handle cyclic bidirectional flows A to B to A Disadvantages Not Excel or table! (Can be an advantage too!) Sometimes flows can looks similar Needs dataset to be in a specific way Can be difficult to interpret compared to Sankey

7 Visualizations Used and Why?
Sankey can be used to show the flow of students from one major to another Chord Diagram can help to show the sequence of courses that students take and grades they get. Since students can take courses out of order, there can be bidirectional flows. Chord diagram works better than Sankey for course data

8 Overall Workflow Choose the type of visualization
Prepare the dataset using Veera Write Code in R Write the necessary code for Shiny Publish to Shiny Server

9 Objectives Visualize the number of number of students migrating from one major to another

10 Data Preparation in Veera
Migration between Majors Current Format One row per student per term enrolled It has major, ids, college and other details Required Format Source Target Flow Source Major (Original Major of students) Target Major (Majors they transferred into) Number of Students

11 Data Preparation in Veera
Migration between Majors Final Format Final Format with years

12 Data Preparation in Veera - Migration between Majors
Fall 2010 Counts the number of students Fall 2011

13 Data Preparation in Veera - Migration between Majors
Calculates the diff bet the first term and the enrolled term to find the year Job called multiple time for different years Writes to two different outputs

14 Visualizations in R Shiny - Migration between Majors
All Years - all_years/ All Years by Source College - all_years_by_source_college/ Each Year by Source College each_year__by_source_college/

15 Objectives Visualize the sequence of courses that students take in a specific program Visualize the effect of grades in one course, on the courses that follow that

16 Data Preparation in Veera
Course Sequence Current Format One row per student per term per course registered It has major, course, grades, ids, college and other details Required Format Source Target Flow Source Course Target Course Number of Students

17 Data Preparation in Veera – Course Sequence
Rest of the data cleanup

18 Data Preparation in Veera
Course Sequence Final Format

19 Visualizations in R Shiny – Course Sequence
Chord Diagram

20 Documentation & Workflow in Veera

21 Documentation & Workflow in Veera
Setup the users and connections Development in Sandbox Folder Create Documentation Move the job to QA folder and share for QA Prod &Test and share with user and schedule Moveup folder share with Admin Admin Users

22 Documentation & Workflow in Veera
Setup the users and connections Admin sets up the connections and the folders. Windows active directory accounts are not used All are veera based accounts maintained by admins Connections are setup using service accounts

23 Documentation & Workflow in Veera
Development Always use shared drives for input excel or text files and use full path. Create separate input and output folders for Veera Use Query nodes and put in SQL with a basic select *, rather than importing tables Save the job under sandbox folder. Outputs The output of test jobs should always write to Test schema using the service account They should all have _test in the file names or table names Create veera outputs under shared drive by creating appropriate folders

24 Documentation & Workflow in Veera
 Add numbers to your nodes by right clicking and renaming it. This helps to debug the steps when a job fails.  Add notes to each node by right clicking it

25 Documentation & Workflow in Veera
Source control veera job using Github by exporting the jobs

26 Documentation & Workflow in Veera
Document everything in a web page or a link Each step number in the veera job will have the details in the documentation page.

27 Documentation & Workflow in Veera
When you are inside a job, click on job-edit notes and add a detailed note about the job and its documentation link

28 Documentation & Workflow in Veera
QA and Prod Moveup Once the job is done, copy it to QA folder and share it with some other user to review it Once QA is done, copy the job to the Moveup folder and share it with admin Admin will check the outputs, make sure it is changed to prod and then schedule the jobs. The admin will share the jobs in Prod and Test folder with the user, for running or debugging in the future Test jobs will be properly labeled by admins

29 Includes only First Time UG cohorts
Pulls in courses and grades for each term ID = 12345 Good Grade Fall 2010 Spring 2011 Math 1 Math 2 Chem 1 Chem 2 Pulls in courses and grades taken after the current course in the terms greater than the current term

30 Thank you! Questions/Comments?
Bharathwaj Vijayakumar, Senior Business Intelligence Analyst, Rowan University #RIConf2019


Download ppt "Using Veera with R and Shiny to Build Complex Visualizations"

Similar presentations


Ads by Google