Week 1: All present. Discussed class website and having to push back our schedule for design, how we will host our DB and actual front end, and how our data will be generated.

• Week 4: All present. DB is running and we are able to insert data into tables from anywhere with an internet connection.
  A small python program was created to bring data over from excel to expedite the process.
  Frontend – Working on frontend mock up for group. Meeting with frontend 2/8 to go over layout and start code.
  ML – Currently narrowing down machine learning implementation using tensorflow by referencing similar real world examples found online. The most promising example so far is forecasting energy usage from data. Continuing to look for further examples that more closely reflect the nature of our project.
• Week 6: All present. DB task has been shifted for a data creation tool. A general, randomly generated item and subsequent quantity program has been made. With concerns on how well it will model real world needs, a simulated model of a store is now in the works. This will include things such as weather, holidays, and an average item use that is modified by the two “events”. This will go day by day and modify average item use for the day with weather conditions and holidays. Certain holidays see an increase in traffic and certain weather events may see a decrease in traffic. However, there are weather events that bring more people into the stores. These include snow, sleet, etc. that could shut down roads for a time. Once an item gets close to being depleted, it will need to consider shipping time and order before it runs out. Emergency orders could become a thing with this system.
  Frontend – Design is done. Current task is implementation.
  ML – I was able to narrow our ML pipeline to Tensorflow 2 library and the Keras framework. We will have to separate our data by store first then export the data from the database to a csv file. Using we can use the timestamp on the order date to create time series data and then Tensorflow/Keras will handle the prediction from that. It is a little more complicated then some examples I’ve seen so far because we have to actually predict two separate values; the item and the quantity. I am not sure if it would be more efficient to do both predictions at once or if two separate prediction processes will be needed.
• Week 8: All present.
  Kyle update: I am pretty confident I have the ML strategy figured out for the data that will be generated. Once we have the generation tool finalized I will be able to start coding the ML. For now I am shifting my focus to preparing our report/presentation that will primarily focus on the data generation tool. I can also assist on any tweaks to the tool if needed.
  Data gen tool: For the most part this is done. Additions that are being added is weather adjustments to traffic and a module to export the generated data to excel for the ML model to use. Could be interesting to see if any changes occur in the data. The next addition would be to insert more variability that cannot be accounted in weather or holiday data.
  Web Page update:
• Week 10:All present.
  Maggie: Currently teaming up with George and Abby to work on the front end of the project. Trying to get the project to run locally on my device and setting up pages stemming from main page
  Kyle:ML Update 4/5 — Created a Google Colab to run the ML Python code. So far I have set up the proper imports needed for the project and worked on the program outline. Currently, I am working on importing the data via csv importing using keras based on some examples I’ve seen. Next, I will store the imported data into tensors and organize it into time series data to prepare it for being processed by Tensorflow.
  Josh: Currently I am working on a visual interface for the data generator. Additions to the data generators calculations may also happen.
  George: Current task is using django to route the buttons on the main page and also setting up log in credentials
  Abby: Working on login functionality.
• Week 12: All present
  ML-Machine model is complete and working on a pipe-lined implementation to go from data generator to predictive ordering generated by the ML.
  Webpage – Integration with DB to host files generated by the data generator and predictive planner.