IS428 2016-17 Term1 Assign3 Heng Yi Teng Mabel

From Visual Analytics for Business Intelligence
Revision as of 20:54, 22 October 2016 by Mabel.heng.2012 (talk | contribs)
Jump to navigation Jump to search

Data Cleaning

Building Data

I wanted to create a variable Floor/Zone so I used the Stack function in JMP.

Stack mhyt.PNG


Result:

After stack mhyt.PNG


Next, I used the Text to Columns function to create 3 new variables: Floor/Zone, Floor and Zone.

New variables floor-zone.PNG


I did this transformation on all the building data, including the Hazium sensor data, and got the combined data file below.

Building data mhyt.PNG


Finally, I used Stack to combine all the Building Elements into one variable. A sample from the final data file is as shown below. Final building data mhyt.PNG


Sensor Data

I needed (x,y) coordinates for mapping on Tableau. The fixed sensor data, proxOut-MC2.csv, did not have coordinates. To assign them with coordinates, I uploaded the ProxZones floor plans onto Tableau (see section on Creating a Map for Sensor Data for more detail) and used the Annotate Point function to retrieve approximate coordinates for each zone.

Annotate point.png

Annotate point tooltip.png


For each zone I took 2 (x,y) coordinates. Next, I matched the coordinates to the fixed sensor data based on the variables Floor and Zone.

2 (x,y) coordinates.PNG

Matched fixed sensor data with coordinates.PNG


I then randomly assigned each record to a coordinate within the respective zones.

Randomizing coordinates.PNG

Randomized coordinates.PNG


Next I used JMP to Join the fixed sensor and mobile sensor datasets based on timestamp, proxid and floor. I did this because I wanted to check if there were instances in which both the fixed and mobile sensors recorded the same activity.

Checking fixed and mobile sensor data for duplicates.PNG

Both fixed and mobile sensors detected.PNG


In these cases, I chose to keep the mobile sensor data as it already comes with (x,y) coordinates. I then deleted the repeat fixed sensor data. Lastly, I combined all the sensor data into 1 dataset.

Final sensor data.PNG

Visualization Process

Building Data

TBC

Creating a Map for Sensor Data

I first loaded the mobile sensor data, proxMobileOut-MC2.csv, into Tableau. On the Data Source tab, I changed the Geographic Roles for X and Y to Longitude and Latitude respectively.

Assigning geographic roles to x & y mhyt.png


Next, in a new worksheet, I dragged X to Columns and Y to Rows. However, we can see below that the data points are superimposed onto a default world map.

Dragging x to columns and y to rows.PNG


Instead, we want to upload a map of the floor plans for the building. First, I manually cropped the VAST_ProxZones_Fx.jpg files and removed the white spaces, leaving only the floor plan. The result is something like this:

ProxZones F1 crop mhyt.jpg


Next on Tableau, I uploaded the images under Map. As there are 3 floors, I had to upload 3 different floor plan images and set different conditions for them.

Maps backgroundimages.png

Maps backgroundimages settings1.png

Maps backgroundimages settings2.png

Conditions backgroundimages.png


Lastly, I used the variable Floor as a filter to toggle between views of the map. This yielded the following:

Map view mhyt.PNG

Your Comments, Please!