Inspiration
We were inspired by the power of big data, and specifically by how twitter feed analysis can provide insight into political, demographical and social dynamics. With a handful of geolocated tweets and query specific search we can understand how a representative set of a community think and behave in different scales. But why is this information relevant? We realise that having live input on a specific query and location provides essential information for business trying to receive input on their performance, advertising companies, and even newspapers and news companies, internet providers etc.
What it does
What we set out to do is a service that performs sentiment analysis on a query specific, geographic specific search, allowing for the user to obtain real time data on the search. Namely, our product performs three functions. First we allow for the user to input a query, a location and a radius for the search, obtaining the sentiment results. Second, we also provide the trending queries list for a specific geographic location, and its correspondent sentiment analysis. We built a html web environment but were short off applying the functionality because we had code in python and java and connecting the back end to the front end was very hard (considering that we were just learning python and also first Hackathon).
How I built it
We used Twitter's developer API and free services to obtain the tweets and analyse the data, using java and python (the latter for the sentiment analysis). For the sentiment analysis we used a Vader model with the NLTK python API, which has been trained with a large dataset in the English language and assigns positive/negative and neutral values to each word and set of words. For the web application we coded the front end in HTML and JavaScript. We tried our best to use Fetch to connect the front end JavaScript to Flask running in the back end, however we couldn't successfully accomplish that. However the separate modules work.
Challenges I ran into
Learning twitter API, learning python, back end/front end functionality, html design.
Accomplishments that I'm proud of
Functional code that reads from live data and provides the output.
What I learned
How hard coding in real life is.
