@NEURO_NUTS

DATA VISUALIZATION
CHOOSING THE RIGHT
VISUALIZATION FOR YOUR
DATA. (PART-I)
 02

Table of contents
Pivot Tables
Line chart
Area Plots
Scatter Plots
Box Plots
Bar Charts
 03

The Dataset

Because discussing the various types of charts is tedious, let's look

at a simple example. Our data set is a simple look at how people
live in a city on a daily basis. Each row is a record of various events
that occurred in a specific district, and we have 1000 of such data
points. These occurrences can be either categorical or numerical.
Our goal is to examine and interpret this data set in order to better
understand the patterns and dynamics of the people.
 04

Lets first get to know more about the transportations used. The
question goeas like this "In the 6 districts, how frequently are the
different modes of transport used?"
Now this is a question that needs to be answered (and can be
answered) with a simple numerical representation.
Though this can be simply represented as table, for larger and more
complex datasets, we generally visualize the heatmaps using different
techniques like a heatmaps or tree maps, which we will discuss about
in Part-2
 05

Area Plots

Since we checked the most used mode of transport in each district, we

want to make a further analysis on the choice of mode of transport.
Hence, we see how frequently each mode of transport is used in each
day. For these kinds of trends, we use what is called as an Area Plot.
Since the number of people using each transport changes over time,
we need an ideal chart that can represent volume changes over time
across multiple categories. This can be achieved using an Area plot.
Area charts combine line and bar charts to show how one or more
groups' numeric values change over time, we will learn about these
two charts soon in the series.
 06

Bar charts

After determining the most popular mode of transport, you

examine the average commute time of each district in all modes to
determine their popularity. Bar charts show the average commute
time per district.
Bar charts, also called column charts, display numbers for each
categorical feature level using bars. Charts show levels and values
on opposite sides. Each category has one bar whose length is its
value. We use a stacked bar chart to show an entire district's
average commute time by mode.
 07

Line Charts Temperature (Normalized)

1

0.75

0.5

0.25

0
Early Morning Dawn Morning Late morning Noon Afternoon Evening Dusk Night Midnight

Now that we got a clarity regarding the mode of transport that

most people use in each district, let us go into a different direction,
and focus on the temperature fluctuations of the city. How does
the temperature of the city change throughout the day?
To answer these questions that are related to trends, we generally
use Line Charts.
Line charts show data as "markers" connected by straight lines.
Line charts are powerful for showing trends over time, forecast
data, and anomalies within and across data series.
 08

Scatter Plots

You found hospital admissions after checking temperatures and

transports. Poor air quality may increase hospital admissions, but
we are unsure. We have the air quality index in our data, so we can
simply check for a correlation.
So, how do you check for these correlations? You use Scatter
Plots. This plot shows the relationship between two different
features by using dots. The values for each data point are shown
by where each dot is on the horizontal and vertical axes.
These plots show relationships between numeric variables, by
using which we can make our conclusion related to the influence of
air quality. (Which the figure shows as no relation.)
 09

Box Plots

Since air quality has no effect on hospital admissions, you think crime rate
might be the reason. Hence to draw that result, you start by identifying
high-crime districts using crime data.
This is more complicated than the correlation or trend checks because the
crime rate can be very high due to a gang fight or absurdly low due to the
strict officer's inspection. You need a plot to visualise all the outlier
conditions, mean, and median crime rates in a single plot. Box plots shine in
such cases.
A box plot shows how one or more numbers are spread out by using boxes
and lines. The line in the middle is the median value, and the edges of the
boxes show the range for the middle 50%. Lines coming out of each box
show the range of the remaining data, and dots outside of the lines show
data related to the outliers.
 10

Coming next in Part-2 ...

Histograms
Network charts
Funnel charts
Heatmaps
Radar plots
Bubble charts
Waterfall charts
Tree charts

Along with a better storyline unlike the last minute story like this ones...
 @NEURO_NUTS

DONT FORGET TO
LIKE AND SHARE THIS
POST. STAY TUNED TO
NEURONUTS FOR
UPDATES ON PART-2!