My Experience as a Data Scientist Intern

In this article, I will share my experience during my internship at Bukit Vista Hospitality Management as a Data Scientist. My initial expectation of being a Data Scientist Intern was that I will be analyzing a lot of data, doing some statistical stuff on it, and making a model, but turns out it was more than that. Data Scientist is more about how we make things more effortless for others. For me, the parable of Data Scientist is like when we want to go from country A to B, the data scientist was the person to advise us what is the most efficient vehicle to reach country B, etc.

As we know Data Science consists of three things maths, computer science, and business. For me, the hardest thing to build was business skills. Because we can’t learn business skills easily without any practice. Unlike math and coding skills, business skills need to be trained by experience in real-world problems. That’s why during this internship, I focused on learning a lot about how to integrate business skills with my maths and coding.

My Journey

In my first two weeks, I learned the fundamental of the tools used in Bukit Vista, the culture, and the flow of the work there. From the first two weeks, I learn how to manage my time, create an OKR for the work that I did, and blend in with the employees. From the second week until the end of the first month, I was working on my first project, Property Classification based on review. Since every guest review on each property has a certain point that represents the property itself. From processing the data from the database to building a decent model for classification. The goal of my project was to classify the property so it can be used for property branding by the Partnership team. By using the guest reviews, I extracted the feature importance of each review and create a classification model to classify the property into 4 classifications is it a comforting property or is it a luxury one. After finishing the code, my first thought was my project was done, but it was not. I have to deliver my model to the client, which was the Partnership team, so it can be adopted as a useful tool by them. Although the presentation was successful, I realize that I still lack business skills. From there, I tried to be more active in scrums and meetings to give some insights and ideas to exercise my business skills to be sharper. I did not expect my insights and ideas were useful to the team. I become more open to giving insights and ideas to others because of that.

In my second month at Bukit Vista, I was given a project to deploy a calculation model for occupancy. I learned how to create an API and deploy a model so it can be scalable enough to be used, efficient, and effective. By creating an endpoint that returns the calculated occupancy I was able to overcome it as desired.I also created a visualization of the database with SQL. I tried by learning from the internet. Luckily, my mentor also helped me out. I managed to create the visualization according to what the team needed. At first, I created the relational algebra and tried it out until I found the most optimal and effective one. After that, I just create the query based on my relational algebra.

In my last month, I learn how to make innovations. Innovation is a new idea that can solve the current problem that can bring an impact on others. I manage to create a topic modelling for the review. With the past guest review on each property, we will know what are the property’s best features eg. the architecture, the big pool, the view, etc. The goal was to help the marketing team to promote the selling point feature of the property. The first approach that I tried was using the word count of all the reviews on each property. But the result was inaccurate. Then I tried to use LDA (Latent Dirichlet Allocation), by using LDA I get a much more accurate result. So, I went with that model and was able to bring the tools to adoption by the marketing team.

In my the same month, I also get a chance to discuss Revenue forecasting with Data Scientist at Airbnb. We discuss what was the best way to deal with time forecasting and how COVID-19 affects the time series model. And we also discuss GAM the model which is the extension of GLM using the GLMM model. It was a fun discussion, I feel honoured that I can discuss it with fellow data scientists at Airbnb.

What I learned from my internship as Data Scientist

1. Data Scientists is not always about creating a model.

   Data scientists especially in startups will require you to work on diverse projects and be involved in the business process as well.

2. Speak Up

   Always actively give your thoughts when internship, so you know if your thoughts are right or wrong. And whenever you face difficulties never be afraid to seek guidance because that is why you took an internship in the first place to learn.

3. What is Data Science?

   Remember, that the most important thing for a data scientist is to create an impact as much as you can on the company using data. And always go for a simple solutions that have a big impact.

4. Learn as much as you can

   Try every field that you think will suit you and the more passionate you are about that field, you will give it effort willingly. In startups. And when you already knew what field your lack, try to improve it.

5. Ad Hoc Request

   When you are on your internship there will be a lot of ad hoc requests. And sometimes the request will not relate to your current position but it’s okay it’s also time to learn.

Conclusion

From my internship, I improved a lot on how to improve my skills related to Data Science. I knew what I learned at Bukit Vista will help me on my journey to become a Data Scientist. Apart from the technical skills, I learned a lot about soft skills and I did a lot of fun with my teammate during that internship. I’m very grateful for the opportunity that Bukit Vista gave me.

It was a great experience that I would never forget.