The problem Pycoders23 solves

Our platform provides a one stop solution for the customers of an E-commerce site as well as the brands. We provide unique recommendations which have a very low false positive recommendation. It does not blindly look for products purchased by the user in history, instead it uses a smart algorithm which takes the categories of purchased products int account for recommending products.(For instance it makes no sense to recommend laptops because customer purchased laptop because it is very unlikely that he will buy another one. But on the other hand we need to recommend customer similar books based on previously purchased books. Our product handles such situations very well.) On the other hand we are addresssing the issue of lack of skilled data scientists and their affordability. We provide Tableau like features for those who are not skilled statistician and automate the process of data visualisatiojn for those. We provide automated customer clustering to segment customers into groups based on the Singular Valued Decompositions and plot them on a graph. We optimally find the number of clusters and provide unique insights for the clusters based on the data. Using these insights the brands can create marketing strategies to target those groups whose effectiveness can be analysed using our dashboard. We also analyse the growth revenue of the brand over certain period. Apart from this we provide a very easy to use GUI where user can put an excel file with his customer data and we provide the same visualisations for them as well.

Challenges we ran into

Eliminating the false positives in the recommendations was a major task given the varied activities of customers. Visualisations had there own set of challenges since the data was not well structured.