Inspiration

Love for learning

What it does

Our solution to the problem statement given comes in the form of a mobile app that has login functionality (to ensure that only the authorized stakeholders have access) as well as the desired dashboard that shows the projected inventory needed for future orders. Future projections will be ahead by at least the order lead time (1 mth?), allowing for stakeholders to be able to respond accordingly and adjust their orders for the various phone models. The dashboard will show this projection in the form of a dotted line graph, which will be updated each day to reflect the actual sales made as a solid line (the dotted line still remains, in-case there is a deviation from the expected trend). Current inventory will also be available in the dashboard menu, alongside other functionalities to update the inventory values manually/automatically, to add new products or remove obsolete ones (EOL), and to account for inventory manually/automatically (including displaying of orders that were made, either by self or by other stakeholders).

For the prediction models, we used multivariate linear regression to train our machine learning model. The data we used for our model is from the mock data sets provided in the tool kit. We identified some key factors such as the launch dates of the mobile phones, whether the weeks had public holidays, mobile phone launch dates, as well as the dates where campaigns were held. We believe that these factors will influence the total number of sales of the different mobile phones, hence we included the data of these factors in our model, alongside past trends. After training our model with the datasets provided, we tested the accuracy of our model and used it to predict the sales of different mobile phone models.

How we built it

We split into different roles, one to design the app using Figma, the other to code the front-end using react framework, two to work on the statistics using scikit-learn to come up with a model prediction algorithm.

Challenges we ran into

Daily life obstacles that created time constraints. Struggled with the complexity of the model prediction

Accomplishments that we're proud of

Finished a working prototype within 48 hours despite the time constraint.

What we learned

Perseverance, teamwork.

What's next for Team 77_InventCast

Another hackathon!

Share this project:

Updates

Submission history