CAPSTONE PROJECTS
PRAXIS BUSINESS SCHOOL, BANGALORE
BATCH OF JULY 2019
Smart Tray: A Restaurant Analytics Dashboard
Everything you need to know about your restaurant
A new restaurant, operating for the last 6-7 months, wanted to understand how they are doing in the market, like, their sales, customer preference, customer values, etc. In this project, we use techniques like Market Basket Analysis, Customer profiling and segmentation techniques to design an insightful and interactive dashboard for this restaurant so that they can analyze their performance and their customer preferences.
ARUS: Attendance Record Update System
Convert handwritten attendance records in digital form
Most schools and colleges record attendances in attendance registers. These registers go to the admin office where the sheets are scanned to keep digital records. The attendance records are manually updated in a spreadsheet or a tracker. In this project, we aim to build a system which can automate the manual updating process and save hours of time.
Automated Attendance Tracking System
Just one click and the attendance is taken...
“Just "one" click and the classroom attendance is taken”. Be it school or college, don't you think the first five minutes of every class is wasted in taking the attendance manually? The objective of this project is to design a system that will be able to record the attendance of a class using Face detection and recognition algorithms.
One Tap Updater: An OCR App.
Update handwritten forms in spreadsheets
Many organizations like banks and insurance companies give customers empty forms to submit their applications. The applicants fill in the applications and submit. But eventually, the data from these forms are needed to be fed into the machine. Our product can help these organizations in updating the contents of these forms in spreadsheets with just a single ‘tap’. In this project, we considered a particular case and shown how this system can work for updating the submitted college registration forms.
Preferencia: A Restaurant Recommendation App
Cross-border restaurant recommendation using customer reviews
Given my preferred restaurant in Dubai can my restaurant recommendation application tell me a few similar restaurants in Sydney? When we talk about similar restaurants we are probably not just talking about food. Other factors like ambience, environment, music, etc. are considered important by the customers. For example, a customer may find a restaurant similar because they have similar ambience and environment, irrespective of the cuisine they provide. Therefore, we considered these factors as important. This project was done in collaboration with ‘Krowd Analytics’, where we used restaurant review data to identify similarity.
Review Ranker: Get the Most Useful Reviews
A review ranking app that saves your time
Potential customers read online reviews to get ideas about the products which they would like to purchase. With a large number of people posting their opinions about the products in the form of reviews it is becoming increasingly difficult for a reader to filter the relevant reviews. However, this task would have been way simpler if we could rank these reviews based on their utility. This is the focus of the project. This will save a lot of time for the review readers because they will be able to derive a good idea about the product by reading fewer but useful reviews.
KindRed: Get the Most Similar Questions
We help you find the most similar questions asked in a Q&A forum.
Different websites like Quora, Stack Overflow, etc. welcomes users to post their questions. These questions are then pushed to the users who may find it relevant to answer them. When the questions are answered the query raiser gets a notification and the question answer pairs are available for the other users to refer. However, with the increasing number of questions being asked in such forums the amount of duplicate (or similar) questions are increasing. In this project we attempt to identify if the question posted is similar to any of the questions which are already answered so than the algorithm can push those question-answer pairs to the users immediately so that the users may refer to them until their questions get answered.
COVID-19 Q&A Bot: Know Facts about Coronavirus
A closed domain question and answering system
Question Answering (QA) system is concerned with building systems that can automatically answer a correct answer to the questions asked by humans in natural language using either a pre-structured database or a collection of natural language documents. Closed-domain Question Answering (CDQA) system deals with questions under a specific domain (for example, medicine or automotive maintenance), and can exploit domain-specific knowledge frequently formalized in ontologies. The ongoing COVID-19 pandemic has brought the world to a standstill and it’s still turning out to be a nightmare across the globe. Our objective is to design a question answering system with the help of Natural Language Processing (NLP) techniques that can answer the natural language queries related to Coronavirus.