1
00:00:00,400 --> 00:00:01,900
Welcome back.

2
00:00:01,900 --> 00:00:04,060
We've seen that we can
pass in series data,

3
00:00:04,060 --> 00:00:08,488
such as lists, into our figure object,
to ultimately generate our plot.

4
00:00:08,488 --> 00:00:12,650
We can pass in other sources of data
as well, such as pics, numpy arrays or

5
00:00:12,650 --> 00:00:14,200
Panda's data frames.

6
00:00:14,200 --> 00:00:17,380
However, Bokeh provides
its own implementation

7
00:00:17,380 --> 00:00:19,980
of series data that we can use,
the column data source.

8
00:00:21,120 --> 00:00:26,010
Column data source is a table like
data object which maps column names

9
00:00:26,010 --> 00:00:27,690
to sequences or arrays of data.

10
00:00:28,710 --> 00:00:32,130
It takes in a data series such
as a Python dictionary and

11
00:00:32,130 --> 00:00:34,190
maps the key to the sequence of values.

12
00:00:35,270 --> 00:00:38,990
In fact, behind the scenes,
when we pass in x equals list one,

13
00:00:38,990 --> 00:00:43,750
two and three, bokeh is making
a column data source with a name of x

14
00:00:43,750 --> 00:00:46,050
with a mapping of values as one,
two, and three.

15
00:00:47,200 --> 00:00:49,260
It also allows for access in the values.

16
00:00:49,260 --> 00:00:53,160
Inside the data, much like one
would with a Python dictionary and

17
00:00:53,160 --> 00:00:55,630
easily accepts pandas
data frames as values.

18
00:00:56,890 --> 00:01:00,510
In upcoming videos we'll also look at some
additional features of having our data

19
00:01:00,510 --> 00:01:03,910
in a column data source,
then make using it in our Bokeh projects

20
00:01:03,910 --> 00:01:07,960
more useful than sticking with
pandas data frames or numpy arrays.

21
00:01:07,960 --> 00:01:11,630
We can add in data to the data source for
hover tool tips and

22
00:01:11,630 --> 00:01:14,908
even use the column data source for
linked visualizations.

23
00:01:14,908 --> 00:01:17,835
Again, we'll be taking
a look at this feature soon.

24
00:01:17,835 --> 00:01:21,370
There's something to note
about column data source.

25
00:01:21,370 --> 00:01:24,890
Like in the previous video,
the columns must all be the same length.

26
00:01:24,890 --> 00:01:28,400
Missing or ragged data is not
permitted within a single data source.

27
00:01:29,460 --> 00:01:32,630
Let's have a look at how we can utilize
column data source to handle some data,

28
00:01:32,630 --> 00:01:35,680
and start to visualize
some actual figures.

29
00:01:35,680 --> 00:01:38,660
To get started,
you want to download the course files and

30
00:01:38,660 --> 00:01:40,230
get them set up in your favorite editor.

31
00:01:41,660 --> 00:01:43,661
I'll head back into PyCharm.

32
00:01:43,661 --> 00:01:47,669
The starter files include the dataset
we'll be using, country-pops.csv,

33
00:01:47,669 --> 00:01:52,470
along with the requirements.txt file which
we already installed in a previous video.

34
00:01:52,470 --> 00:01:58,250
Let's go ahead and
create a new file called stage1-3.py.

35
00:01:58,250 --> 00:02:01,620
Here, we'll want to get
our CSV file imported and

36
00:02:01,620 --> 00:02:03,870
take a look at the format of our data.

37
00:02:03,870 --> 00:02:07,710
Since this file contains information for
all the countries of the world,

38
00:02:07,710 --> 00:02:12,600
let's remind ourselves why we're
exploring data to just five rows and

39
00:02:12,600 --> 00:02:14,000
display the header information as well.

40
00:02:15,110 --> 00:02:18,800
We'll need to import numpy as np,

41
00:02:20,810 --> 00:02:27,310
import pandas as pd, again, that's a pretty
standard naming convention for those.

42
00:02:30,810 --> 00:02:35,846
We're gonna be bringing in
our country-pops.csv file.

43
00:02:41,383 --> 00:02:45,718
And we'll use the pandas.read_csv method.

44
00:02:47,860 --> 00:02:48,729
Bring this in.

45
00:02:51,866 --> 00:02:57,332
And then for data exploration,
we'll create a np.array,

46
00:03:01,339 --> 00:03:04,203
With a header, and

47
00:03:04,203 --> 00:03:09,516
we need to limit our rows here, nrows=5.

48
00:03:14,541 --> 00:03:15,963
And print(countries_array).

49
00:03:20,510 --> 00:03:25,393
And when we run this,
we can see then that in our data,

50
00:03:25,393 --> 00:03:28,550
we have quite a bit of useful information.

51
00:03:28,550 --> 00:03:34,340
The English and German names of
the country, code code, population, etc.

52
00:03:34,340 --> 00:03:36,760
As is often the case with datasets,

53
00:03:36,760 --> 00:03:41,450
there's information included that you
might not necessarily be interested in for

54
00:03:41,450 --> 00:03:44,740
the current project,
such as birth and death rates.

55
00:03:44,740 --> 00:03:49,030
But, it's great to have that information
available for our future explorations.

56
00:03:49,030 --> 00:03:52,550
Now that we have a table of data,
we should put it to use.

57
00:03:52,550 --> 00:03:56,600
Imagine that you have this information in
your favorite spreadsheet application.

58
00:03:56,600 --> 00:03:59,750
Let start with the chart that people
will often generate in Excel.

59
00:03:59,750 --> 00:04:01,520
A bar chart for example.

60
00:04:01,520 --> 00:04:03,624
We'll need to bring in our imports and

61
00:04:03,624 --> 00:04:06,657
in this case import bokeh.charts
to plot a bar chart.

62
00:04:19,320 --> 00:04:23,280
And from bokeh.io, we'll import
our output file and show methods.

63
00:04:25,850 --> 00:04:27,430
Then we define our output file.

64
00:04:27,430 --> 00:04:30,043
Let's call it population.html.

65
00:04:36,742 --> 00:04:40,932
Next, we need to build out our bar chart
by passing in our countries pandas

66
00:04:40,932 --> 00:04:42,560
data frame.

67
00:04:42,560 --> 00:04:45,610
We can do that here at the bottom and
tell it which information to use.

68
00:04:47,940 --> 00:04:49,114
We'll just call it bar_chart.

69
00:05:21,158 --> 00:05:26,315
Next, we need to pass in our bar
chart object into the show method.

70
00:05:28,970 --> 00:05:33,595
show(bar_chart) and then run our script.

71
00:05:38,285 --> 00:05:42,725
Since our n rows is still set to five, we
see the five countries displayed on a bar

72
00:05:42,725 --> 00:05:46,110
chart showing their
representative populations.

73
00:05:46,110 --> 00:05:47,260
Pretty cool.

74
00:05:47,260 --> 00:05:50,320
You'll notice that we are setting
our legend to false so

75
00:05:50,320 --> 00:05:51,900
that it doesn't display.

76
00:05:51,900 --> 00:05:55,860
In a later video, we will take a look
at how to customize our chart legends.

77
00:05:55,860 --> 00:05:59,080
There are a variety of other charts
that can be generated as well, and

78
00:05:59,080 --> 00:06:01,600
I've included links to them
in the teachers notes.

79
00:06:01,600 --> 00:06:04,470
While, pandas data frames work great for
many applications.

80
00:06:04,470 --> 00:06:08,480
Bokeh provides another option that
allows for some cleaner code.

81
00:06:08,480 --> 00:06:12,810
And as we'll find out in a later video,
some great additional features.

82
00:06:12,810 --> 00:06:16,440
To take advantage of this,
we utilize Bokeh's column data source.

83
00:06:17,470 --> 00:06:20,910
This will map the column names
to the sequences of data.

84
00:06:20,910 --> 00:06:24,730
Let's go a bit beyond bar charts,
or pie graphs with their data.

85
00:06:24,730 --> 00:06:28,110
Since we're attempting to provide some
data analysis in our reporting, and

86
00:06:28,110 --> 00:06:32,030
see what, if any, correlation there is
between the country's population, and

87
00:06:32,030 --> 00:06:33,620
life expectancy.

88
00:06:33,620 --> 00:06:38,051
Before we can use it, we need to import
column data source from bokeh.plotting and

89
00:06:38,051 --> 00:06:39,630
pass in our data.

90
00:06:39,630 --> 00:06:44,787
We'll need to do our imports from
Bokeh along with the column data

91
00:06:44,787 --> 00:06:50,906
source import from bokeh.plotting,
we want to import ColumnDataSource.

92
00:06:54,763 --> 00:06:56,080
And figure.

93
00:06:56,080 --> 00:06:58,360
Since countries' a pandas data frame,

94
00:06:58,360 --> 00:07:01,790
we can pass that into
ColumnDataSource as our data source.

95
00:07:01,790 --> 00:07:05,521
And since we don't need the country
array or print function any longer,

96
00:07:05,521 --> 00:07:07,980
we can clean up our code and
delete those lines.

97
00:07:09,820 --> 00:07:13,800
And, while we're at it, let's rename our
output file to something more meaningful.

98
00:07:14,880 --> 00:07:21,320
Since we're examining population
versus life expectancy,

99
00:07:21,320 --> 00:07:26,158
let's use pop-life.html as our file name.

100
00:07:26,158 --> 00:07:28,150
And we can get rid of
the numpy import as well.

101
00:07:29,320 --> 00:07:32,027
And we won't be needing bar any longer.

102
00:07:38,573 --> 00:07:39,073
Great.

103
00:07:40,210 --> 00:07:45,100
We wanna set up our country data.

104
00:07:45,100 --> 00:07:49,640
Just taking our ColumnDataSource,
and we want our countries.

105
00:07:53,630 --> 00:07:56,456
Now, we just need to build
out our figure plot,

106
00:07:56,456 --> 00:07:58,867
similar to what we have done in the past.

107
00:07:58,867 --> 00:08:01,587
But, we'll use circle ellipse here and

108
00:08:01,587 --> 00:08:05,270
not pass in any tool parameters
to accept the default.

109
00:08:06,730 --> 00:08:07,494
Our plot.

110
00:08:13,595 --> 00:08:16,010
And we'll label our x_axis as Population.

111
00:08:21,150 --> 00:08:26,090
Our y_axis is Life Expectancy.

112
00:08:31,140 --> 00:08:35,933
Great, we'll do a plot.circles,

113
00:08:35,933 --> 00:08:40,423
pop in population for our x value.

114
00:08:45,409 --> 00:08:46,549
Life_expectancy.

115
00:08:48,803 --> 00:08:49,762
For y value.

116
00:08:52,130 --> 00:08:55,115
Our source data is our country data.

117
00:08:58,319 --> 00:08:59,940
And we'll make our glyphs 15 points.

118
00:09:01,250 --> 00:09:05,620
That should look roughly familiar, but
what is going on there with our x and

119
00:09:05,620 --> 00:09:06,480
y values.

120
00:09:06,480 --> 00:09:09,990
And what does that source
equals country_data bit?

121
00:09:09,990 --> 00:09:13,320
Well, we are telling our plot
to use our source variable,

122
00:09:13,320 --> 00:09:18,240
our ColumnDataSource information,
as the source of our data.

123
00:09:18,240 --> 00:09:22,010
Our x value is a population
value inside that data set, and

124
00:09:22,010 --> 00:09:27,570
our y values are the values of the
matching Life Expectancy column of data.

125
00:09:27,570 --> 00:09:32,020
Notice that capitalization matters here,
and that we need to match our x and

126
00:09:32,020 --> 00:09:35,600
y values to the correct names
of our data set column headers.

127
00:09:36,630 --> 00:09:40,600
Now, we just need to tell Bokeh to
show us our plot and we'll be all set.

128
00:09:42,990 --> 00:09:45,340
Show plot, and run our script.

129
00:09:51,160 --> 00:09:55,320
Awesome we have a plot showing us
life expectancy versus population for

130
00:09:55,320 --> 00:09:56,950
our five countries.

131
00:09:56,950 --> 00:10:00,050
Great work,
we have seen several things here.

132
00:10:00,050 --> 00:10:04,416
How to utilize pandas data frames as
a source of data and get it to work with

133
00:10:04,416 --> 00:10:09,157
Bokeh column data source, and plot that
data based on specific column names.

134
00:10:09,157 --> 00:10:12,995
It's not very helpful right now though in
terms of data exploration because we can't

135
00:10:12,995 --> 00:10:14,230
tell which glyph is which.

136
00:10:15,870 --> 00:10:19,110
When we come back, we'll look at adding
specific colors to our glyphs and

137
00:10:19,110 --> 00:10:22,210
add legends to our graph to help
better understand our data.