We are a team of International Students who arrived in the US only a few months ago. We chose this project because it will help people get familiar with a new country easily on their fingertips.

What it does

Columbus, the tour guide is the dream one stop shop for any explorer. Are you in the mood to explore the city or do you yearn for some of that delicious local food?, Columbus is at your service!! It can help you out with: 1) Suggesting different things to do like activities, food, etc. 2) Places to visit 3) Suggesting different routes to a place based on mode of transportation the user prefers. 4) Fun facts/tips about a place

How we built it

First, we built the chat bot using Rasa framework. It's incredibly powerful, and is used by developers worldwide to create chatbots and contextual assistants. Then, we imported the useful datasets from the Yelp database and queried them using SQL to get the appropriate results for a user's question which we filter based on location. After that, we transferred these results to our Twitter sentiment analyzer to get the suggestions ranked according to the sentiments of people who are talking about the place . We developed a python code which uses Twitter API to read the tweets for a particular keyword/hashtag and assign it a positive, negative and neutral score using TextBlob and NLTK sentiment analyzers. TextBlob uses a Movies Reviews dataset in which reviews have already been labelled as positive or negative and is trained on a Naive Bayes Classifier. While displaying the ranked suggestions, we also provide users with a fun fact/tip about the particular place using Yelp data. Finally, we created the UI for the chatbot using HTML, CSS and JavaScript.

Challenges we ran into

Some of us were new to Rasa framework and usage of Twitter APIs, Sentiment analysis using TextBlob. So, analyzing and fixing the issues/errors we faced was a bit of a challenge. Also, it was a challenge working with MapBox as it was our first time using it but the result was well worth the effort!!

Accomplishments that we're proud of

We are proud that we were successfully able to help Columbus answer tourism-related questions using Yelp data and rank the suggestions based on popular sentiment of people from Twitter. Having first-hand experienced the problems a person faces in a new city/country, this will help users to get familiar with a new place easily with reliable, trustworthy information present on their fingertips as Columbus leverages the huge Yelp database and the power of real-time Twitter data.

What we learned

We got the wonderful opportunity to learn about Rasa framework, the usage of Twitter APIs and sentiment analyzers to perform real-time analysis on Twitter data. The feeling of accessing social tools used by billions of people and then extracting valuable data to make a persons life easier was heartening. Having first-hand experience in the problems we tackled, it is invigorating to visualize the massive impact that it is going to have on future travelers.

What's next for Columbus the Tour Guide

We will be adding another useful feature which will help Columbus answer questions related to safety of a place based on the crime rates and other incidents information about that location. Also, using the COVID data, we will give health-safety suggestions to the users like is a particular place the user want to travel to is following COVID safety protocols or not, etc. Also, we are going to implement Speech To Text(STT) and Text To Speech(TTS) translation so that blind people can access this service too.

Built With

Share this project: