thedarktrumpet / 22c196-moteproject

22C196 Project for Fall 2012

Readme for TheDarkTrumpet

About Me

I've been an application developer for since, professionally, 2005 and unprofessionally since the mid 1990s. I've also been heavily involved in Data Analysis and Machine Learning since around 2012. I'm proficient in a number of programming languages, with my primary languages being Python, Golang, C#, Common Lisp, and a few other scripting languages. I hold a Master's Degree in Computer Science with an emphasis in Machine Learning.

A good deal about me is available on MY WEBSITE and MY RESUME

22c196-moteproject's Issues

Create Midterm Presentation

The midterm presentation should include:

1. A basic powerpoint presentation on our project, specifically a reintroduction to our project.
2. Screenshots from our application, showing the status we've made
3. A XLS file showing the plots of signal strengths received from the application in some environment.
   3A) A few graphs showing the division in "room guesses"

Develop a very simple graphical test

Harpreet,

This goes with your task on researching graphical options, but please take the map that Eric sent around and try and figure out a way to show this map within a GUI of some kind, and to be able to draw a position on it. Since this is kinda a prototype, you can simply do this in straight Java, or on the android, your choice.

Create a reformat script for the CSV output

Right now, the CSV format is a bit hard to follow. Thinking a bit about formatting, the following formats may be best:

<MAC_ADDR_1>,Signal_1,Signal_2,...Signal_N
...
<MAC_ADDR_N>...

The graph I'm hoping to see is on the X axis a listing of MAC addresses, then on the Y Axis is a color-coded plot of their strengths. This should show clear divisions for each of the rooms per mac address.

I decided to do this logic in a combination of common lisp, and linux commands. Will document my progress under this task.

Update scroll text view

Eric, please add context about how this isn't being refreshed.

Look into Classification options

Amit,

You seem to have some level of knowledge of data mining principles. If you would, please look into the KNN and the algorithm the professor mentioned during class. I forget what algorithm this is, so please:

1. Email the professor and find out the name of the algorithm (today)
2. Prepare a short 2-slide presentation about that algorithm and KNN (or any other algorithm you believe will be good for this project)

This will be for this Sunday's meeting.

Update Lisp prototype program to output a CSV file of ranges.

Using the Common Lisp version, output the range-information into a CSV file. in the format:

ROOM,MAC_ADDR,RANGE_LOW,RANGE_HIGH

Investigate Graphical Overlay

A graphical overlay will eventually need to be done, that can show a virtual map of the building, and where we are at. This isn't a development task, but create a small presentation about how this can be done.

Multiple CSV Files

When hitting "stop logging", it should flush out to a single CSV file, and close the handler. When starting logging, once again, it should open a brand new CSV file.

Format:
wifi-value.<DATE_TIMESTAMP>.csv

Implement "room level" identification

Either using Weka, lisp, or whatever, come up with a way to identify us being in a particular room. This doesn't have to be within the project framework, but may be worthwhile adding it there, and having a graphical element that supports it (maybe another action that one can go to).

This task is actually quite a few tasks bunched into one. I will split them out if "big enough"

Implement distance metric

This is just to implement the very specific function used to do what you're proposing. Amit may have some questions, since he's actually going to be using this in his task.

Create a reformat script for the CSV output

Right now, the CSV format is a bit hard to follow. Thinking a bit about formatting, the following formats may be best:

<MAC_ADDR_1>,Signal_1,Signal_2,...Signal_N
...
<MAC_ADDR_N>...

The graph I'm hoping to see is on the X axis a listing of MAC addresses, then on the Y Axis is a color-coded plot of their strengths. This should show clear divisions for each of the rooms per mac address.

I decided to do this logic in a combination of common lisp, and linux commands. Will document my progress under this task.

create 2-slide presentation on more precise area in room

Eric,

We already discussed this, but to make things clear, please create a 2-slide presentation on how we'd more accurately define our location in a room.

I do still believe we're a ways from implementing this, but I think talking about it early makes sense.

Create java functions to load the new data into a hash object

The hash should be:

"ROOM_NAME" = { "MAC_ADDR" => [SIGNAL_LOW, SIGNAL_HIGH],
.....
.....}

Gather wifi hotspot meaures

3 rooms on the university. Ideal would be JH. Should be 2 rooms and 1 hallway (connecting the rooms)

This should be 3 CSV file sent to me, 1 for each room.

Reformat CSV File

The CSV File should be formatted as:

TIMESTAMP_1,WIFI_MAC_ADDR_1,WIFI_STR_ADDR1,....,WIFI_MAC_ADDR_N,WIFI_STR_ADDRN

Using distance metric, implement a simple "here's my coordinate"

Amit,

as discussed during the meeting - please take the image of SC, and "mock" out a position on that map. This should be using Eric's implementation of the distance metric (he has the task now).