Is my life boring?

When my life is interesting I take a lot of photos. Lets see if I can plot these photos to see if my life is getting more interesting.

In [1]:
base_dir = "/home/visgean/Dropbox/**/*"
In [2]:
import glob
import os

def get_extension(filename):
    filename, file_extension = os.path.splitext(filename)
    return file_extension.lower()


picture_extensions = ['.jpg', '.jpeg', '.png']
pictures = list(filter(
    lambda f: get_extension(f) in picture_extensions,
    glob.iglob(base_dir, recursive=True)
))
In [3]:
print('# of pics:', len(pictures))
# of pics: 3136
In [4]:
import exifread

with open(pictures[40], 'rb') as f:
    tags = exifread.process_file(f)
    
print(tags.keys())
dict_keys(['MakerNote Tag 0x0001', 'Image Orientation', 'EXIF FlashPixVersion', 'Image YResolution', 'Image XResolution', 'GPS GPSLongitudeRef', 'GPS GPSImgDirectionRef', 'EXIF SensingMethod', 'Thumbnail ResolutionUnit', 'GPS GPSImgDirection', 'JPEGThumbnail', 'MakerNote Tag 0x0006', 'GPS GPSAltitudeRef', 'EXIF ColorSpace', 'EXIF ExposureBiasValue', 'EXIF SceneCaptureType', 'EXIF FocalLength', 'Image ResolutionUnit', 'Image Make', 'EXIF SubSecTimeOriginal', 'EXIF LensSpecification', 'EXIF BrightnessValue', 'Image DateTime', 'EXIF ApertureValue', 'MakerNote Tag 0x0005', 'EXIF FNumber', 'EXIF MeteringMode', 'GPS Tag 0x001F', 'GPS GPSLongitude', 'Thumbnail JPEGInterchangeFormat', 'EXIF LensMake', 'Image Software', 'MakerNote Tag 0x0003', 'EXIF ExposureTime', 'EXIF ShutterSpeedValue', 'Thumbnail Compression', 'MakerNote Tag 0x0008', 'Image ExifOffset', 'EXIF WhiteBalance', 'GPS GPSLatitude', 'EXIF ExifVersion', 'EXIF ExifImageWidth', 'EXIF DateTimeOriginal', 'Image Model', 'GPS GPSDestBearingRef', 'MakerNote Tag 0x0014', 'GPS GPSDestBearing', 'GPS GPSAltitude', 'EXIF SubSecTimeDigitized', 'GPS GPSSpeedRef', 'EXIF ComponentsConfiguration', 'EXIF FocalLengthIn35mmFilm', 'EXIF ExposureMode', 'Thumbnail JPEGInterchangeFormatLength', 'EXIF Flash', 'Image YCbCrPositioning', 'EXIF MakerNote', 'Image GPSInfo', 'GPS GPSSpeed', 'MakerNote Tag 0x0004', 'GPS GPSLatitudeRef', 'EXIF SceneType', 'EXIF ExifImageLength', 'EXIF ISOSpeedRatings', 'GPS GPSDate', 'EXIF LensModel', 'EXIF DateTimeDigitized', 'Thumbnail YResolution', 'MakerNote Tag 0x0007', 'GPS GPSTimeStamp', 'Thumbnail XResolution', 'EXIF ExposureProgram'])
In [5]:
date = tags['Image DateTime']
In [6]:
print(date.values)
2016:10:05 16:16:11
In [7]:
from datetime import datetime
datetime.strptime(str(date.values), '%Y:%m:%d %H:%M:%S')
Out[7]:
datetime.datetime(2016, 10, 5, 16, 16, 11)
In [8]:
def parse_date(exif_info):
    if not exif_info:
        return None    
    date = exif_info.get('Image DateTime')
    if not date:
        return None
    try:
        return datetime.strptime(str(date.values), '%Y:%m:%d %H:%M:%S').date()
    except:
        print(date)
In [9]:
def get_exif(filename):
    try:
        with open(filename, 'rb') as f:
            return exifread.process_file(f)
    except:
        return None
In [10]:
exif_data = list(map(get_exif, pictures)) # i want to iterate through these multiple times
MemoryError at position: 603312351, length: 3945995808
Possibly corrupted field DateTime in Image IFD
Possibly corrupted field DateTime in Image IFD
Possibly corrupted field DateTime in Image IFD
Possibly corrupted field DateTime in Image IFD
Possibly corrupted field DateTime in Image IFD
In [11]:
dates = list(map(parse_date, exif_data))
2013:12:03 19:12:81
In [12]:
from collections import Counter

photos_per_day = Counter(dates)
In [13]:
print('Missing dates', photos_per_day[None])
del photos_per_day[None]
Missing dates 1247
In [14]:
print('Missing dates', photos_per_day[None])
Missing dates 0
In [15]:
from bokeh.plotting import figure, show, output_notebook
output_notebook()
Loading BokehJS ...
In [16]:
time_graph = figure(
    title="Pics over time", 
    background_fill_color="#E8DDCB",
    y_axis_label='# of pics', 
    x_axis_label='Time',
    x_axis_type="datetime"
)

sorted_vals = photos_per_day.most_common()

time_graph.circle([x[0] for x in sorted_vals], [x[1] for x in sorted_vals])
show(time_graph)

As you can see that is a bit messy so lets try to collect it by month

In [17]:
photos_per_month = Counter([datetime(d.year, d.month, 1) for d in dates if d])
In [18]:
time_graph = figure(
    title="Pics per month", 
    background_fill_color="#E8DDCB",
    y_axis_label='# of pics', 
    x_axis_label='Time',
    x_axis_type="datetime"
)

sorted_vals = photos_per_month.most_common()

time_graph.circle([x[0] for x in sorted_vals], [x[1] for x in sorted_vals])
show(time_graph)
In [21]:
print(len(list(filter(None, dates))))
1889

I was only able to parse about 60% of my photos - rest of them comes from facebook or dont have dates at all :/ But its still nice to see that I am taking more and more photos.