[OSM-dev] [GSOC] Suggestion for a Crowd-sourcing Geo-Processing Project

[Saad Qureshi]

Hi,

My name is Saad Qureshi and I am currently a 3rd Year Electrical
Engineering Student at NUST, Pakistan.

I got a chance to read through all the projects that OpenStreetMap is
working on and I would like to recommend an innovative project that me and
my friends have been working on for a few months which, if implemented
completely, might prove very useful for the OpenStreetMap community. The
project is available on github and you can get a basic overview of the
project on main README.md file:

https://github.com/msaadq/aero2

Or Watch the intro Video: https://www.youtube.com/watch?v=RkyMHUqihVs

The project works like this:

Our volunteers or general athletes use our arm-band (BT enabled with MQ-5
sensor) to collect small samples (without them noticing) during their
commute or walking/running sessions from different parts of a city.

We collect that data from our database and relate it to certain properties
of different nodes of a city to a specified resolution (50x50 metres in our
case) and use the samples (of some nodes) and properties (of all nodes) to
predict the samples for the rest of the nodes without physical sampling.

I want to make a formal proposal for this project, but I wanted your
feedback on it first.

This Project has a hardware module, android application module and 2
backend modules.

Hardware:
    - Interfaces the MQ-5 sensor and Bluetooth module with the ARM Cortex
M4 processor for data acquisition over Bluetooth and sending it directly to
the android app (Completed)

Android Application:
    - Uses Mapbox API to visualize the data as heat maps overlay over
google maps for both sampled and resultant data and allows the user to
enable the sensor remotely for collecting samples. (Completed)

Backend:
    - A Python-based backend for collecting the properties of a city using
MapsAPI and saving them in a database for later (Partially Implemented)
    - A Python-based machine learning implementation for using the samples
(of some nodes) and properties (of all nodes) to predict the samples for
the rest of the nodes. (Not Implemented)

This is one of those projects for which I wake up every morning and it
would be really awesome if I get to work on this during GSOC. Your feedback
regarding this project (and its implementation is highly appreciated).

*Please let me know if this project resonates with the interests of *
*OpenStreetMap**, so that I start making a formal proposal.*

Saad Qureshi
www.linkedin.com/in/msaadq
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20160321/ca46d684/attachment.html>