"I worked on three significant data science projects during my internship. First, I was responsible for collecting fashion product data from various brands by using web scraping methods in Python, since the firm's previous way of scraping clothing data was highly customized but time consuming. Therefore, I explored generalized web scraping ways and obtained two big fast fashion brands and nine sustainable brands, which added more than 30,000 products to our dataset, providing sufficient data for subsequent modeling. Second, I contributed to database design and development, storing the data I collected in a structured way. After finding that their past ER diagram was unsuitable for new data, I figured out the new logical way to design the database. Then, I managed the initial data input in AWS using PostgreSQL queries, achieving a solid and valuable database. Last, I improved their current recommendation system and made it more robust. I added the image-based recommendation algorithm and the content-based one and included more similarity metrics to accomplish a higher similarity recommendation. With more clothing data from different brands, Changing Room will have a more comprehensive database, providing solid data support for the recommendation system. The recommendation system is crucial for their first product version because they aim to change consumers’ mindsets in an eco-friendlier way by recommending sustainable alternatives for online consumers, revealing the importance of my three projects. Therefore, my data science internship helped Changing Room parse the information hidden in vast amounts of clothing data, consolidate the data structures, and improve the recommendation system.

My master’s degree in business analytics helped me gain programming skills, business sense, and soft skills. Because I was focused on a data science career path, I registered for many classes in the IEOR department to build up my technical skills, including statistics, machine learning topics, Python coding, cloud analytics, etc. Therefore, I learned many programming skills (including web scraping, website establishment, SQL, data analysis, and modeling) which I applied directly to my summer internship because the data scientist role ranges from gathering and processing data to recommendations using models. Furthermore, the business analytics degree covers business aspects as well. Through my Business School courses (such as Marketing Analytics), I learned many advanced marketing tools and improved my business sense when analyzing a problem. Although my role was not directly marketing related, I could contribute ideas when the meeting required everyone’s input because the startup is now prioritizing user growth. Finally, I picked up some soft skills when networking and attending social events, which boosted my presentation and expression skills. Occasionally. I presented in front of the whole company with people from different teams to give updates. My school experiences provided me with confidence and experience when delivering in front of a big audience. Hence, every aspect of my degree was crucial to my work, and I constantly strengthened these skills by applying them in a real-world setting.

I faced the biggest challenge of generalizing the web scraping methods to gather clothing information from various fashion brands. Due to the fact that their traditional web scraping method was too customized and time consuming, exploring a new approach was important. Plus, the launch of their first version was expected to be in the summer, so the data collection was significant for Changing Room. Accordingly, I did a lot of research online and talked with experienced friends, professors, and my manager to find possibilities for new approaches. Eventually, I managed to figure out the embedded data structure to achieve a generalizable scraping algorithm. Another challenge for me was the image-based recommendation system. Deep learning is not a familiar topic for me in my coursework, but it is widely applied in the image-based recommendation algorithm. Because their original recommendation system is heavily based on the description of clothing items, missing and messy text is always a problem. Before adding the image-based algorithm, I self-studied online and applied a model that best fit our company. The challenges helped me grow, and the process of finding ways to solve them was the most precious experience for me.

My takeaways are twofold. First, I bolstered my programming skills to a great extent, including the flexible application of web scraping methods, the logical clarification of database structures, and the optimization of modeling for the recommendation system. Using what I learned from school and strengthening these skills in practice is a virtuous circle, but applying them to real life problems is not easy because, unlike classes, there are many surprising difficulties encountered during the process. The spirit of exploring new ways is an essential takeaway from my internship because I can benefit from this ability for my entire life. Furthermore, I enhanced my soft skills in interpersonal communication. As a data scientist intern, I need to communicate with the environmental research team about storing weights of different clothing parts in the database and the software engineering team about deploying back-end recommendation systems to the front-end. So knowing how to communicate with technical and non-technical people is vital, especially for non-technical people because they may be shy to ask questions when hearing a technical term. Briefly explaining some technical terms can help them better understand the context.

My goal is to stick to the data career path and combine my qualitative and quantitative experiences to have more impactful output based on analytics. Within the responsibility of this internship, I strengthened my coding skills through engaging in data collection, database design, and recommendation system improvements. Also, with some qualitative analysis that I had this summer, I understood consumer needs and wisely designed the technical solutions. Through this internship, I developed a deeper understanding of users and a higher technical level of achievement, which will significantly help me in my future career path."