[WEEK 2 — Wi-Fi Based Indoor Positioning]
Team Members: Burak Emre Ozer, Huzeyfe Kocabas
In the second week of our Machine Learning Term Project which is Wi-Fi Based Indoor Positioning, we’re diving deep into the dataset.
The UJIIndoorLoc database covers three buildings of Universitat Jaume I. This data set is focused on WLAN fingerprint positioning technologies (also known as WiFi Fingerprinting).
The first 520 features contain the received signal strength indicator value (RSSI) of wireless access points (WAPs) for an instance at a location. RSSI values are represented in a negative form, where the closer the value is to 0, the stronger the received signal is.
We can see that all the reference point location information is found in the last 9 features which contain additional location information along with user and phone information.
Explore the data
import pandas as pddataset = pd.read_csv("train.csv")
dataset = dataset[dataset.columns[-13:]]
dataset.head()
After we transform the dataset into the data frame, we create a 3d-scatter plot based on the location points using 3 different columns which are longitude, latitude, and floor.
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')xs = longitude # dataset['LONGITUDE']
ys = latitude # dataset['LATITUDE']
zs = floor # dataset['FLOOR']
ax.scatter(xs, ys, zs, alpha=0.5, s=10,marker='o')plt.title("Location points with respect to the floor")
plt.show()
In the following weeks, we’re going to deal with several algorithms (eg. k-nearest neighbors, random forest, decision tree) to build predictive models.
