def prepareDataKMeans(lSegments):
for segment in lSegments:
segment['distance'] = distance.getDistances(segment)
segment['velocity'] = distance.getVelocities(segment)
segment['speedClass'] = initSpeedClass(segment)
segment['numSC'] = initSpeedClass(segment)
return lSegments
def importJson(filepath, addColumns=True):
# Loading data
raw = pd.io.json.read_json(filepath)
df = raw['locations'].apply(pd.Series)
# Create latitude and longitude columns
df['latitude'] = df['latitudeE7'] * 0.0000001
df['longitude'] = df['longitudeE7'] * 0.0000001
# Clean up columns
columns = ["timestampMs", "latitude", "longitude"]
for col in list(df):
if col not in columns:
del df[col]
# Add date column in format 'dd-mm-YY'
dates = []
for timestamp in df['timestampMs']:
dates.append(
datetime.datetime.fromtimestamp(int(timestamp) /
1000).strftime('%d-%m-%Y'))
df['date'] = dates
# Add time column in format 'HH:MM:SS'
time = []
for row in df['timestampMs']:
time.append(
datetime.datetime.fromtimestamp(int(row) /
1000).strftime('%H:%M:%S'))
df['time'] = time
if not addColumns:
return df
else:
# Add delay column seconds
delay = []
delay.append(0)
for i in range(df['timestampMs'].size - 1):
delay.append(
int(int(df['timestampMs'][i]) - int(df['timestampMs'][i + 1]))
/ 1000)
df['delay'] = delay
# Add distance, velocity and acceleration
df['distance'] = distance.getDistances(df)
df['velocity'] = distance.getVelocities(df)
df['acceleration'] = distance.getAccelerations(df)
return df
def filterBySegment(input_data, limite): data = input_data.copy() lat_filtered=[0]*len(data) long_filtered=[0]*len(data) data = delay_segment_dataframe(data, limit=limite) segment_count = max(data['segment']) segment_indexes = [0] j=0 for i in range(data['segment'].size): if data['segment'][i]!=j: j+=1 segment_indexes.append(i) segment_indexes.append(data['segment'].size+1) for k in range(segment_count): distances = data["distance"][segment_indexes[k]:segment_indexes[k+1]] dist_moy = np.mean(distances) if dist_moy<20 and segment_indexes[k+1]-segment_indexes[k]>8: data = medianFilterSegment(data, 3, segment_indexes[k], segment_indexes[k+1]) elif dist_moy<50 and segment_indexes[k+1]-segment_indexes[k]>6: data = medianFilterSegment(data, 2, segment_indexes[k], segment_indexes[k+1]) elif dist_moy<150 and segment_indexes[k+1]-segment_indexes[k]>6: data = meanFilterSegment(data, 2, segment_indexes[k], segment_indexes[k+1]) # Add distance, velocity and acceleration data['distance'] = distance.getDistances(data) data['velocity'] = distance.getVelocities(data) data['acceleration'] = distance.getAccelerations(data) return data