The client who approached us had noticed that the overall portfolio performance of US Senate/Congress members frequently and sometimes significantly outperformed the average US stock market performance. However, none of the current platforms could give a consolidated and clear picture of trades done by these individuals.

These were the challenges that we had to solve:

• Analyze and consolidate the trade data from several different trading platforms

• Normalize the data consolidated and create a clear structure in which we would display the data

• Analyze not only the retrospective profitability of the trade

• Use multiple machine learning models to forecast the possible profitability of each trade even at the very moment of the purchase

When creating product like this, the precision and accuracy is everything. We made sure we were as precise as possible for every aspect of the product. Here is what we did:

• We delve into the tiniest details for each feature that was needed and talked with client on all different approaches that we could take

• We created rough wireframes to give the shape to the ideas that were pre-agreed to make sure that we interpreted the scope the way the client had envisioned it

• We created the PoC of the product to test out the accuracy of algorithms we used

• We developed the product in chunks, allowing very frequent interactions with the client. Allowing aligned, spot-on delivery

We built the following things for the product:

➤ Robust database

That records different trades make on multiple trading platforms through public APIs available.

➤ Interactive dashboard

That allows use to get bird-eye-view on all recent trades that particular Senate/Congress members made, details of the operation, retrospective and forecasted profitability.

➤ Admin Panel

Allowing the users to handpick the Senate/Congress members they want to track.

➤ 3 different lightweight Machine Learning models

That we used for predicting trade profitability from the moment the trade was made

This lean, feedback-first approach helped ensure that every new feature addressed an actual user need, not just an assumption.

We were able to deliver the entire MVP within 2 weeks! And fit within our initial estimates and budget that the client had in mind. The results were great:

✔️ Product is slick and easy to use

✔️ The dashboard enabled users to get the full oversight of the data within seconds

✔️ The models we built performed the way the client expected

We never got to measure the results of our ML models, given that the product was intended to be for the private use, but we got great feedback, and we are proud of the quality that we delivered.