PROJECTS

The government of Spain was considering an expansion of its renewable infrastructure investments. As such, they required information on the trends and patterns of the countries renewable sources and fossil fuel energy generation. We worked on modeling a load shortfall prediction algorithm. This is essential for the operation and planning of utility companies.

We were tasked with analyzing the supplied data, identify potential errors in the data, and clean the existing data set. Determining if any additional features could be added to enrich  the existing data set, and then build a model that is capable of forecasting the three hourly demand shortfalls.

Additionally, determining what features were most important in the model's prediction decision, and explain the inner working of the model to a blended audience.

Lastly, we deployed and API version of our model using Flask API.

Many companies are built around lessening one's environmental impact or carbon footprint. These offer products and services that are environmentally friendly and sustainable, inline with their values and ideals. Further, they would like to determine how people perceive climate change and whether or not they believe it is a real threat. This would add to their research efforts in gauging how their product/service may be received.

The objective of this project was to create a machine learning model that is able to classify whether or not a person believes in climate change, based on their novel tweet data. As we know Twitter data is a powerful source of information on a wide list of topics. For the purpose of this project, climate change data from tweets were analyzed to determine whether they believe in climate change or not.

The tweets analyzed were divided into four classes: 

• News: Tweets linked to factual news about climate change.

• Pro: Tweets that support the belief of man-made climate change.

• Neutral: Tweets that neither support nor refuse beliefs in climate-change.

• Anti: Tweets that do not support the belief in man-made climate change.

Our model was hosted on a AWS EC2 instance. Lastly, our team deployed a user-friendly web-based application, by merging our machine learning model with Streamlit. 

South Africa is a multicultural society that is characterized by its rich linguistic diversity. Language is an indispensable tool that can be used to deepen and also contribute to the social, cultural, intellectual, economics and political life of the South African society.

The country is multilingual with 11 official languages, each of which is guaranteed equal status. Most South Africans are multilingual and able to speak at least two or more of the official languages.

For this project, I have taken text from South Africa's 11 official languages and identify which language the text is in. This is an example of NLP (Natural Language Processing), the task of determining the natural language that a piece of text is written in.

In today's technology driven world, recommender systems are socially and economically critical to ensure that individuals can make optimized choices surrounding the content they engage with with on a daily basis. One application where this is especially true is movie recommendations; where intelligent algorithms can help viewers find great titles from tens of thousands of options.

I worked collaboratively to design, build, and deploy a Movie Recommender System. Such recommendation systems are beneficial for organizations that collect data from large amounts of customers and wish to effectively provide the best suggestions possible. 

We modeled both a content-based and collaborative recommender system. Both of which were hosted on an EC2 instance. We then mounted an S3 bucket to the instance, which stored our data.