InspectYour Gadget

Inspiration

Technology is rapidly developing every single day, and new laptops are constantly being added to the market. Often, customers struggle to find the best laptops for their needs, due to the difficulty of finding accurate and helpful information in the face of obstacles like minimal digital fluency and sponsored content masquerading as genuine advice. We endeavor to bridge this divide by offering a bias-free, user-centric platform to empower individuals to make informed decisions in their pursuit of the ideal laptop.

What it does

InspectYour Gadget is a digital platform designed to demystify the process of finding the laptop that best suits the user’s preferences and lifestyle. Hosted on a user-friendly website, our product features a short survey that allows users to consider and rank their priorities in selecting a laptop to purchase. Based on the survey responses and product data fetched from popular tech seller sites like BestBuy, our backend processes use the OneHotEncoder and K-nearest neighbor machine learning algorithms to identify the top 3 product listings that will give the user what they are looking for in a laptop.

Through this personalized and dynamic approach, InspectYour Gadget is able to communicate real-time product data to users, eliminating the guesswork from the purchasing process and encouraging educated purchases.

How we built it

For our development, we used Jupyter Notebook on Intel Developer Cloud, which proved to be an invaluable resource in providing the additional computing power we needed to run our data analysis and machine learning algorithms.

Using the Beautiful Soup python library, we first scrape live data from top tech seller websites, using BestBuy as a case study, on the various laptops and computers available on the market. Then, this data is vectorized via the OneHotEncoder ML technique and fed into a machine learning model that employs the K-nearest neighbors (KNN) algorithm to determine the products with specs most closely matching the user’s inputted preferences.

To host our product, we made our website using HTML, CSS, JavaScript, and Django, opting for a high-contrast theme. For a smooth user experience, all dropdown menus and buttons change color upon hovering and have various other effects to highlight the most important information on the page.

Challenges we ran into

None of our team had used Django in the past, so there was a steep learning curve as we were beginning our project and learning how to host our website. We also struggled with processing the web-scraped data into a categorical format that could be analyzed by our ML model. However, the bulk of our time was spent integrating the backend work with the user-facing website, communicating back and forth as our product collected user input from the website, fed it to our data processor, and outputted the results back at the end.

Accomplishments that we're proud of

We are proud to have successfully completed a functional project with a working backend and a user-friendly frontend design. We are particularly proud of our web-scraping and data analysis functionality, as well as of our growth over these past 24 hours, having learned how to incorporate many new and useful development tools in a short amount of time .

What we learned

We learned how to work with a wide array of tools and technologies because of this project. When getting data for our model, we learned how to employ various libraries and APIs in web-scraping. Specifically, we used Beautiful Soup, Python, and Regular Expressions to scrape data from Best Buy’s website and vectorize the data. During this project, we also learned how to use Django and how to effectively debug challenges that we ran into, such as dependency issues and git merge conflicts.

What's next for InspectYour Gadget

In the future, we plan to implement more sophisticated backend processes to improve the preference-matching process and provide even more accurate, personalized laptop recommendations for users. Specifically, we hope to add more preference options, such as durability, battery life, compatibility with accessories and other devices, and current deals on the market. We will also spend more time on improving the user interface with more accessibility features, such as support for voice readers and light/dark modes.