April 23rd, 2020 Team Zoom Meeting
AI engine backend successfully delivers an output in a format that is readable by the frontend. Beginning to try out PostgreSQL connection to see if that functionality can be added.
April 22nd, 2020 Team Zoom Meeting
AI engine backend successfully takes in csv data delivered by the frontend via JSON, and runs everything correctly.
April 21st, 2020 Team Zoom Meeting
Received final feedback on how to finish off the project to hand to client and present a demo. The only things left are to make sure the backend can read the request, and send a response the frontend understands. If time permits, added PostGreSQL database connection functionality will be added to the demo.
April 17th, 2020 Team Zoom Meeting
Discussed how to wrap up the project. Almost got the connection with the frontend to backend figured out. Once this is complete, then the project should be presentation/client-ready
April 17th, 2020 Extra Coach Meeting
Met with Coach to work through connecting javascript and Django
April 14th, 2020 Coach Meeting
Discussed progress with connecting AI engine inputs with javascript
April 12th, 2020 Team Zoom Meeting
Worked on and discussed translating SQL queries to MapReduce functions in Javascript to connect frontend inputs for AI engine to read in properly
April 10th, 2020 Extra Coach Meeting
Met with Coach to work through connecting javascript and Django
April 8th, 2020 Team Zoom Meeting
Added additional functionality to AI engine to allow for testing with different ML models such as XGBoost, multi logistic regression, KNN, and a simple neural network.
April 7th, 2020 Coach Meeting
Made decision to move forward with connecting frontend application to backend AI engine
March 31st, 2020 Coach Meeting
Presented working AI engine that takes in data from PostgreSQL database and transforms data into a form that logistic regression model understands, and gives a list of predicted outputs.
March 27th, 2020 Team Zoom Meeting
Most of preprocessing and pipeline in place. Slight modifications may need to be made, but nothing major. Some testing with a logistic regression model was made to see how everything would work.
March 25th, 2020 Team Zoom Meeting
Began work on data preprocessing.
March 24th, 2020 Coach Meeting
Discussed what is to be completed this week. Deliverable to complete by next coach meeting is to have data preprocessing and pipeline in place. Will show a 5-10 minute demo of the working pipeline next week. The preprocessing/pipeline is set up to begin model implementation and accuracy testing for next week.
March 22nd, 2020 Team Zoom Meeting
Further discussion of hosting Postgresql for cloud, created a backup in case this didn’t work. Created a script that took our data from Google Sheets and uploaded it into Postgresql database.
March 20th, 2020 Team Zoom Meeting
Talked about what remote work would look like for project completion. Began working out how to host Postgresql for everyone to have access to.
March 1st, 2020 Team Meeting
Worked on project presentation.
February 28th, 2020 Team Meeting
Constructed a plan for how to build the data pipeline for AI model. Most of the work will be proper feature selection, preprocessing the data, and creating relevant metrics for the input to train the data with. Likely models will be either Knn or Logistic Regression.
February 25th, 2020 Coach Meeting
Discussed project logistics based on the resources available given to us by the client. Determined that focusing on our own inputs/data types for the AI model would be best, along with data that we create for testing purposes.
February 21st, 2020 Client Meeting
Discussed with the client the kind of data we want to gather to construct the AI model. Gained access to Heroku where PostGresSQL is located. The client is going to look at what they can do on their end to start tracking and generating data, meanwhile, we are going to look into what kinds of models we would likely use along with trying to generate some fake data to build some of the data pipelines with.
February 18th, 2020 Coach Meeting
Decided our focus for the week was to gain access to the data that our client currently has.
February 16th, 2020 Team Meeting
Put together architecture design based on what we learned at the Client meeting.
February 14th, 2020 Client Meeting
Talked with our client’s software developer lead, and was able to gain information on what systems they have in place. The backend framework is implemented using Django and hosted on Heroku. Django communicates what content should be delivered to Roku by sending JSON files. Roku implements Brightscript code that creates frontend user interaction.
February 11th, 2020 Coach Meeting
Discussed with coach about client expectations. We may need to choose between connecting front end client data to the database, or creating the AI model in a vacuum. This is dependent on what information we get in our next client meeting, and when the code gets delivered to us.
February 9th, 2020 Team Meeting
Put together specifications document and the platform selection.
February 5th, 2020 Client Meeting
Discussed with the client the general idea of what is wanted for the AI project, which is to implement a system in which the consumer is recommended content based on what they are interested in. We also discussed gaining access to the app source code and database, which will be provided as soon as correspondence with the more tech-savvy client has been established via Slack.
February 4th, 2020 Coach Meeting
Discussed with our coach what our goals were for the client meeting. We agreed that letting the client talk as much about the product as possible would be useful, as well as doing our best to gain access to the app code.
February 2nd, 2020 Team Meeting
Discussed potential user stories and project concepts before meeting with the client.
January 28th, 2020 Coach Meeting
Met with our coach and discussed the feasibility of certain parts of the project. Planned to meet with the client to discuss the project further.