Phase 3 Dialogue Reflection

#

Phase III of the project was a whole other beast and was by far the most time-consuming phase. Perhaps a day after phase II ended, as a team we consulted with Dr. Fontenot due to confusion on how to connect personas based on emission tags. We then changed our ER diagram and made tags a new strong entity with 3 many to many relationships between the personas. The SQL DDL was updated accordingly. I did some work by myself throughout the week based on the wireframe from phase II. I made the initial landing page and figured out page navigation for some pages. Later that week I added NGO enterprise matching functionality with two get API routes which matched NGOs with enterprises based on shared tags. This was my first time working with the API and I was pretty confused about all the things I needed to change with the actual blueprint, making the blueprint in rest_entry.py and calling the route through the API URL. However, it was a great learning experience and I eventually figured it out which made writing future routes a lot easier. Next, I added the Enterprise compare tool which was another get request. This feature shows the logged-in enterprise’s emissions, their country, and the average of all enterprise’s emissions within that country. In my free Sunday in Luxembourg, I tried to add an NGO info section page that would have the NGO user fill out the information to be posted into the table (POST request). This feature took me extremely long for no reason. I thought I had the logistics right with collecting and sending the data but I kept getting an error. I had a back-and-forth Slack conversation with Dr. Fontenot about it and decided to leave it till Monday. It turned out my api url was slightly wrong causing the whole thing to not work. I eventually managed to make the successful POST request and that feature was completed. On the day before phase III was due we met as a group and had a productive conversation about where this app was going to go. We discussed with Dr. Fontenot that the user persona’s logging in should already have survey/info data filled in. However, I had already created the NGO info questionnaire. Thus, we decided to make the info questionnaire a PUT request to edit information and the surveys would also be PUT requests if users wanted to retake their survey. We decided to leave the survey off until the due date because, on the Data Science side, we still needed to finalize what questions we were going to ask (if they actually impacted emissions). I instead worked on adding history pages for each persona (displays survey result history). On the day of the due date, we slaved away. We still didn’t know what survey questions to ask due to the data science side not being done. So I changed the NGO POST route to a PUT route instead given the NGO user is already logged in. After continuing to work on setting up pages, we got news from Michael that there wasn’t a strong correlation between many of the initial factors we anticipated and emissions. In fact, only two features he found were strongly correlated meaning we only could have two emissions questions. This was obviously very disappointing, but there was little we could do. Finishing up the night, we implemented a bunch of the features and I helped Anjola create a POST, and DELETE request for changing and adding NGO tags. I also hard-coded country and emission data into our database, as well as mockarooed some of the entities. I assisted Aahil in the integration of the ML model and the finishing touches. Then, edits were made to the blog and we called it a night.