Stock Dice Mobile (a successor of its desktop version - Stock Dice) is a mobile application for stock market data collection, analysis and prediction (to some extent).
The idea behind Stock Dice Mobile is to collect stock market data over time, (over an year or so) and use that data for market trend analysis and prediction. The prediction was not very accurate at the very moment of demonstration, but it was a proof of concept, with Scrapper, Prediction Engine, API and thin client all in a working model.
When starting out on this project (back in December 2013) as my final year research project for graduation, I was new to PhoneJS (http://phonejs.devexpress.com) . I decided to develop a thin client using this framework as my final year project.
From a developer's perspective the most interesting part for me was to perform analysis by using ChartJS (http://chartjs.devexpress.com) charts (line, pie etc) in the mobile application, showing the right amount of data in the small screen realstate that was available to me at that moment.
And hey! with little effort, that worked on almost every phone :)
It allows the user have certain level of personalization by offering favorite companies and favorite sectors, to confine the analysis and prediction and display more relevant results or data.
Personalization |
The prediction accuracy can be improved using different algorithms (i was using Linear Regression at that moment).
Prediction Accuracy Chart |
I presented that project on Open House Job fair 2014 at University of Engineering and Technology Taxila, and was awarded First Position.
Here are a few more snapshots of the project.
Top Volume Sectors |
Forecast Global |
Forecast Details |
The snapshots and the material included in this post is the sole property of the owner , Basit Munir, and cannot be reproduced in any form. Violations should be reported at (basitmunir147[at]gmail[dot]com).
No comments:
Post a Comment