At our Engineering Colleges, each Faculty/Teaching staff is evaluated on how their students have fared in the Subjects that they have taught. The semester results for hundreds of Engineering colleges under RTU University is sent via email in a unstructured csv file. There is a dedicated team of admins in each college, who convert the excel manually and then initiate Faculty Analysis. This csv contains results of 50000+ students each semester across different colleges. And each college does the similar repetitive work, each Semester; and since last few years.
What it does
We have built a Faculty Scorecard Cloud Bot - which is a zero footprint bot for any of the Colleges who are interested, to just send an email from an authorized email ID, with raw RTU results sheet. Once the mail is received, bot starts it job, parses the excel, generate scorecard report for each Faculty and also aggregates them at a College level; and emails back the generated PDFs back to the authorized email ID. All this happens in 5 just minutes.
How we built it
We have used the following flow and technologies to develop the bot Receive/parse Email attachments --> Store in a smart DB structure (so that history is built over time for reactive /predictive analytics) --> Define PDF report templates for Faculty Score Card and College Score Card --> Generate PDFs with extracted/processed/stored data be using templates --> email back the generated PDFs to the authorized email
Technologies Used UI PATH MySQL pdfgeneratorapi.com
Challenges we ran into
UIPath Excel Parsing with unstructured csv content Identifying all Human interpretations of csv data No capabilities in UIPath on Reporting/templates mgmt; and hence switch to external PDF template engine
Accomplishments that we're proud of
- The bot eliminates the following effort . 100+ admins (each college has an administration dept) who takes 5-10 days every 3 months (semester exams), i.e. more than 3000 days of work is fully automated
- And this bot will not complain of monotonous work - and probably will never ask for compensation increase; and will never publish erroneous results which could jeopardize faculty's career
What we learned
- With UIPath / RPA engines, it is seriously easy to automate many such tasks that are repetitively done in Academia and Corporates
- The Bot kick starts the journey for more such Bots - as it enables people to think what all can be automated and make their life better.
- UI Path needs reporting/templating capability so that any ops effort that results into a report, can be used through its own inbuilt functions. We are thinking of adding a integrated library to make this happen
What's next for Faculty-Scorecard-Bot
- Enroll all Colleges
- Provide historical reports ( with no PII data) that can show case relative strength of each Faculty/College over two dimensions : 1. Individual historical improvements 2. College and/or Faculty Comparison over time with all colleges under that University