BizzIn

Inspiration

Smaller businesses struggle with making decisions on where to open a storefront in a new area due to limited access to information and a way to properly and efficiently compare locations. This is exacerbated by limited finances to pay for proper research on market entry in a new area. Our solution simplifies the research process and allows for democratization of information and access to resources for smaller businesses with limited means.

What it does

We are taking in data points that businesses looking to enter a new area would need to consider (such as foot traffic, competition in the area, and market size) and analyzing how well those fit the business's needs and assigning a score to each possible location, along with a description of why they fit the business's needs. Possible locations are found based on retail locations for sale in the area and can be compared with one another to find the best possible fit.

How we built it

The data points in consideration are within datasets (Foursquare, Google Maps, OpenAI) that are fed into a machine learning model (Sci-Kit Random Forest Models) to develop an algorithm to calculate ratings for each location which are then followed by reasoning for the rating (Generated by OpenAI). Possible locations are found using OpenAI to search for For-Sale retail locations in the area, and the web application was created using Next.Js and Tailwind CSS for Front-end and MongoDB for Back-end to store user data and preferences.

Challenges we ran into

With the Google Maps Dataset being removed, our largest challenge was piecing together the data points that dataset provided us on our own, through adding together various other datasets that provided parts of the same information in order to train our machine learning model. This was challenging in not only being time consuming (cleaning and pasting the data), but also in thinking of creative ways to replace what was provided by the Google Maps Dataset and making it difficult to envision the rest of the project without the main functionality of our website.

Accomplishments that we're proud of

We believe that our endurance and perseverance made this possible and are what we are the most proud of. In order to make this project come alive, we had to overcome challenge after challenge that almost made this idea seem like it would be impossible to pursue in terms of complexity and obstacles with dataset cooperation. Our tireless efforts to find creative solutions and perseverance to overcome obstacles were our greatest assets in producing something we believe would solve problems many businesses face.

What we learned

What became clear to us as the Datathon progressed further and further was the importance of prioritization of tasks. Many additional features that we desired to add bogged us down in productivity since we were too detailed in additions that weren't relevant to creating a minimum viable product, leaving us in a bigger time crunch for more integral features. We will take this into account for our next Datathon experience and prioritize a working project while only keeping additional features in mind for later.

What's next for BizzIn

We plan on expanding to businesses that are not only looking to enter a market, but are already established in a market looking to evaluate how well they are doing and in relation to competitors to optimize their strategy. This will also include trends so businesses can visualize whether they are doing better or worse compared to the past to inform better decision making and evaluation, like whether something new they added isn't working as well as expected, or if a certain feature they have needs to be updated. We are also planning on adding more data points in consideration for market entries such as local infrastructure (roads, parking, utility), addition of complementary businesses analysis in competitive landscape, customer demographics, labor force availability, and real estate costs. Additionally, our machine learning model will be trained on a significantly larger dataset for improved accuracy.

Built With

• chatgpt
• gemini
• google-maps
• mongodb
• nextjs
• scikit-learn