1)
Project summary:
Using beta and logistical regressions to answer the simple question: "should retail investors invest in IPOs?"

Goal:
Use historical data to aid investment decisions made by retail investors, in the highly debated (and partly misunderstood) concept of IPOs. Is getting in as early as one can actually the best decision?

Approach:
The statistical anaysis includes data of IPOs from 1998-2019, and takes into account the post hoc performance of companies after they went public, historical performance of the market (S&P500), performance of alternative investments, performance of US Treasuries as a benchmark, and incorporates external factors such as the dotcom bubble early 2000s and the financial crisis of 2008. Post hoc effects of COVID 19 were NOT taken into consideration.

• Data used was a combination of publicly available data and data from Harvard Business School Publishing.

Outcome:
The model used 75% of the data as a training sample, and tested accuracy on the rest 25% test sample. Acoording to the results of random test samples, the model has a 78% accuracy of predicting IPO performances over the following 5 years, specifically for the tech and manufacturing industries.

2)
Project Summary:
Using the programming language R and associated libraries to develop a pricing model for Uber (or any similar ride sharing company)

Goal:
Help new ride sharing services develop a suitable pricing model, one of the toughest business decisions in this field, due to the dynamic nature of all dependant variables and inputs.

Approach:
The model studied the current situation of Uber, based on predicted Uber rides depending on time of day, weather, traffic, no. of regular customers active, and even accounting for inflation. The model automatically adjusts ride price, with the goal of maximising profit for the company (such as increase price in typically rainy months, where people tend to prefer to use ride sharing, however the increase should not be so much as to dissuade new customers). The model used (and is accurate in predicting) ride sharing servies only for Munich, Frankfurt and London.

• Data used was a combination of publicly available data and data from Harvard Business School Publishing.

Outcome:
Based on a training and test sample, pricing models developed using this model would have helped increase Uber profits by 3-5%, and would be of paramount importance to any new ride sharing services operating in metropilation cities.