Ascend

Dawgs: Dylan Stockard Ece Oz

Challenge Statement: Develop an automated solution for the restaurant American Son (owned by Chef Tim Ma) that will read, categorize, and summarize reviews into a report.

Solution: Chef Tim had the problem that reading reviews could be frustrating, unproductive, and he holds personal bias as to how the reviews are interpreted. To avoid the stress of manually assessing reviews, we created a program that can scrape OpenTable - an online reservation and review website for restaurants such as Chef Tim's American Son - to read in the reviews left by users. We then analyzed reviews as positive or negative, and categorized these reviews into 5 categories:

• Brunch: American Son's brunch menu.
• Dinner: American Son's dinner menu.
• Food: The general quality and taste of the food.
• Ambience: The mood and feel of the restaurant setting.
• Service: The quality of service provided by the restaurant. After categorizing the reviews, we calculated valuable statistics and auto-generated a PDF that visualized the customer reviews so that Chef Tim can know at a glance what his customers and did and didn't enjoy with their meal at American Son.

We also listed reviews that were deemed useful based on character length and category in the PDF so that if Chef Tim wanted to read the reviews anyway, he could read reviews concerning exactly what category he was interested in, and avoid reviews that provide no constructive criticism.

Currently, we run the program to produce the PDF report on our end, and then upload it to a google drive that Chef Tim can access after being notified that a new PDF is created. This way he has no responsibilities of his own to create the report, and just has to retrieve from google drive.

Other programs that do a similar task require their own rating service to gather and analyze the reviews. That is, they can't use OpenTable reviews, even though OpenTable is far more convenient for customers since it also provides a platform for booking restaurants. Our program can collect more data by using the reviews from OpenTable.

Moving forward, we have three main objectives to scale our program that are absolutely achievable given time. They are as follows:

• Expand our system to apply to other restaurants. This can be done by adjusting the URL of our web scraper.
• Incorporate reviews from other sources (i.e. Yelp). Each site may have its own way of attaining information, but largely can also be done by expanding the URL range of the web scraper
• Migrate to a web application. This will just take time to develop the UI/UX of the app, but would allow for users to get review information whenever they want in a more interactive format.

A Review for your Reviews is a solution for restaurant owners who are caught up in their own reviews and are ready for an analytical approach to listening to their customers.

Built With

• jupyter-notebook
• python