The application works on the premise that you have data, and that you have specific questions, which you think the data can answer. So you begin by asking the question, and selecting relevant parts of the data, which then, gets transformed into Customizable Visualizations.
At this point you have the ability to create a Dashboard, based out of those questions, that you can store locally on your system Or share* with other users.
Application Flow
Workspace is the thing that binds together, a set of Datasources, related Questions & Dashboards.
You would typically start with providing the data.
We currently support Excel based files, Tabular data & JSON data.
Having uploaded the data, you would move to the Editor, which would be empty to begin with.
Click on Add Question, to start your Data Discovery journey.
Start with the Question text, choose the related Datasource, Chart type, fill in the necessary inputs in the Selections View & click on Process.
The datasource would be parsed, based on the inputs provided, and if everything goes well, would have the selected data. This would also show the Chart View & the Data View sections.
The Data View allows you to see the data that the, selections have picked up, and also the Visualization with the chart, selected earlier.
Having added a question, you can choose to include it in a Dashboard, and then navigate to the Dashboard, to view all the Charts together.
This view allows you to resize & reorder the charts.
Save Workspace
Each workspace is saved to the inBrowser database, and can be obtained in a file format, which can be shared with other users.
Editor
This is where you would spend majority of your time.
It lets you create, update and delete the Questions & comprises of the Questions View Area, Selections View, Chart View and the Data View
Questions View Area
Lists out all the questions, that have been added as part of this Workspace. It allows to Edit, Copy and Delete each question, along with the ability to decide whether this question would be part of the dashboard or not.
Clicking on one of the questions, opens up the Edit area, which would allow you to Select an appropriate chart, change the selections, View & Plot the Data, and Customize the PlotWIP
Chart selector
Select a chart from our set of supported charts. We currently support Vertical & Horizontal Stacked & Grouped Bar, Line, Area, Pie, Table and Hierarchy Charts.
Selection Area
The selection area, updates itself based on the Datasource & Chart type selected. This allows you to specify the columns and the rows that you would want to pick up for processing.
Since we work with raw data, instead of a Point & Click approach, we would need you to specify the columns, rows and the value groups that you would be interested in. And to that end, we have a very intuitive instruction set.
Instructions
Sheet NameOnly applicable for Excel Sheets.
We currently support processing only 1 sheet per question
This is the sheet, that would be used as the basis for the Column & Row selectors.
Column SelectorFor Excel Sheets.
The general structure would be [-]<Column Cell Alphabet><Row Number>[_<Column Cell Alphabet><Row Number>], where [ ] indicates optional.
Single Column
Select Single Column via a combination of <Column Cell Alphabet><Row Number> Eg. B2, C2
Multiple Columns
Select Multiple Columns by giving the start location, and the end location, separated by an underscore ( _ ) Eg. E1_AI1
Exclude columns
If you wish to Exclude certain columns, use the hyphen ( - ) before the selection. Eg. Say you are selecting a range of columns, & wish to exclude a couple of them in between -E2, -K2
Combining selectors
Multiple selectors, would be separated by comma ( , ). ( The order does not matter )
Example :
E1_AI1, -J1_K1, B2
This will pick up columns E to AI, recording the heading in the 1st row, would reject the columns J & K, and also pick column B, but record its name from the 2nd Row, and with result in a total of 30 columns.
Column SelectorFor Table & Raw Data.
Since, the column names are known from beforehand, so the dropdown selectors allow for selecting one or more columns. Based on the Mapping Type however, these selectors might be restrained to one
Row Selector
The general structure for all statements, would be [-]<Column Cell Alphabet><Row Number>[_<Column Cell Alphabet><Row Number>], where [ ] indicates optional.
Single Row
Select Single Rows by giving the row number Eg. 4, 5
Multiple Rows
Select Multiple Rows by giving the start location, and the end location, separated by an underscore ( _ ) Eg. 3_65
Multiple selections
Provide Multiple selections by comma separating the individual selections Eg. 4, 5, 6_65 . ( The order does not matter )
Exclude Rows
If you wish to Exclude certain rows, use the hyphen ( - ) before the selection. Eg. Excluding row 54 as -54
Example :
4, 5, 6_65, -54, -44_48
This will select rows from the third sheet, 4 to 65, except row 44, 45, 46, 47, 48 & 54.
Data Mapping
Data mapping options would vary based on the chart selected, ( might not be available for some ) but roughly there are four major categories [ e.g. for Bar charts ]
Multiple Columns, Multiple Rows
Column names would be on the X-axis, as individual bins. Values for a single column would be summed up across the rows, based on the Data Type & Groups provided
Single Column, Multiple Rows
Distinct values from the column would be on the X-axis as bins. Occurence of these values would be summed up, and plotted on the Y axis.
Single Column, Single Column
Distinct values from the X-axis column, would be on the X-axis as bins. Corresponding entries on the 2nd column would be summed up, based on the Data Type & Groups provided.
Single Column, Multiple Columns
Allows for grouping & stacking. Distinct values from the X-axis column, would be on the X-axis as bins. Corresponding entries on the other columns would be summed, & stacked up.
Chart View & Data View
Based on the selections, the application would parse the data, and get the relevant data. This would be viewable in a Table format. The same would also be modified into a chart specific format, and be viewable as a visualization.
Couple of parameters, whose values are being tracked, across a period of time, maybe a week or a month, measured daily Or over a period of year, averaged over a week or month.
2. Employee attendance, over a month, measured daily.
Employee
D1
D2
D3
D4
D5
D6
D7
D8
D9
D10
Emp1
Emp2
Emp3
Emp4
Emp5
Bar Chart
Days would be on the X-axis
The data might take on a set of possible values ( A, P, a, p, Leave etc. ) i.e. Categorical
Plotted as is, the unique values across each day would be plotted as Stacks.
The selections give the capability to define bins across values, For eg - map A, a and leave as A & P and p as P. This would then have only two stacks per day.
Selections on the Editor, gives you the capability to view related data ( View Related Data ), on clicking of each bar.
3. Say, you are a trucking company, you have a fleet of some trucks, have their trips data, T1 to T10, and the average milage of the trucks on these trips.
Truck
T1
T2
T3
T4
T5
T6
T7
T8
T9
T10
TR1
TR2
TR3
TR4
Bar Chart
Trips i.e. T1 to T10 would be on the X-Axis as individual bins.
The data, would be numerical.
Now, you could either choose to plot data for each truck separately, which then would create a Stacked Bar chart, Or using the Numerical Selector, you can choose to break the data into groups, and then apply one of the operators sum, min, max, average or count, to each of these groups.
Selections on the Editor, gives you the capability to view related data ( View Related Data ), on clicking of each bar.
You have a single column, that can take a bunch of categorical values Or is a numerical value, and has these values over a range of rows.
1. Measurment of a couple of parameters, say Diabetes Type & BMI, for a set of people. Now you would like to break these values across categories and get a count of how many people fall in each category.
Person
BMI
Diabetes Type
P1
P2
P3
P4
P5
Bar Chart
If you are looking to plot BMI, then the categories created via the Numerical Selector would be on the X-Axis.
If you were to plot Diabetes Type, then the unique values, would be on the X-Axis.
The count for each of these categories, would be plotted on the Y axis.
Selections on the Editor, gives you the capability to view related data ( View Related Data ), on clicking of each bar.
When you have a set of values from a single column, and you would like to find the related range of values on another column. Most often this will result in Stacked Bar Charts.
1. Think of a typical college semester, for a particular discipline. A single semester has 6 odd subjects, and each student, would score a certain grade in each subject. So if you are looking to visualize subjects, and the distribution of grades over the subjects, You would choose this mapping.
Student
Subject
Grade
S1
S1
G1
S1
S2
G1
S1
S3
G2
S2
S1
G2
S2
S2
G1
S2
S3
G3
Bar Chart
Having chosen Subject as the X-Axis, you would have bars, corresponding to each Subject. 3 in this case
The Grades being categorical, would be summed up, based on their unique values, and would appear as stack for each subject.
If it were actual marks, instead of grades, then the Numerical Selector would help you break these into groups and the count, across those groups, would be plotted.
Selections on the Editor, gives you the capability to view related data ( View Related Data ), say students, on clicking of each bar.
This could be thought of as the opposite of Multiple Columns, Multiple Rows, since the multiple columns here would be on the Y-Axis, instead of the X-Axis.
You would use this, When you have a set of values from a single column, and have related values tracked over a period of time. Most often this will result in Grouped Stacked Bar Charts.
Let's work through an automobile giant, Hawa Inc . With operations in 8 countries around the world, the company offers 4 models and is looking to track their sales, across regions over the quarters.
Region
Vehicle
Q1
Q2
Q3
Q4
R1
V1
R1
V2
R2
V1
R2
V2
R3
V1
R3
V2
Bar Chart
Vehicles would be on the X-axis.
All the rows corresponding to a particular vehicle, would be picked up, & the values for these rows across the columns Q1 to Q4, would be added up, resulting in grouped bars, one for each column.
If the values in columns were categorical Or numerical values are broken into groups, via the Numerical Selector, then each bar for each column, would be further broken into stacks.
If however you are looking to sum up all the values, across all the quarters, to get a consolidated value for each vehicle, you can use the Sum Up option on the Editor.
Selections on the Editor, gives you the capability to view related data ( View Related Data ), on clicking of each bar.
View By would allow you to view the same data on an individual Region basis.
Chart Types
The application currently supports
Vertical Bar Chart
Horizontal Bar Chart
Line Chart
Area Chart
Pie Chart
Hierarchy or Sunburst
Table Chart
Scenario, Selections & Processing
Scenario
Selections
Processing
MCMR - User has selected multiple columns. The data for these columns across a set of rows is Categorical.
The data needs to be treated as Case Insensitive, which simply means P and p are the same.
As the values are read for a particular column, across the rows, and a count of these is being created, the values will be converted into lowercase and matched.
All the columns should be treated as individual, with no combining together.
All columns would be plotted on the X-Axis
All the unique values across rows should be plotted separately.
Each column then would have a bunch of keys and associated values, resulting in Stacked Bar Charts.
Two or more X-axis columns need to be combined together, as one and given a separate name.
Once data has been collected for all columns, these values i.e. key:value pairs would need to be combined across these columns that should be combined together.
Two or more unique values need to be combined together, as one and given a separate name.
For each of the column, once we have the set of key value pairs, we would look at all of them, and then combine the count corresponding to these keys.
Bar Chart
Mapping
Columns
Chart Type
MCMR
All columns would be on the XAxis
Each column would have a bunch of keys and associated values, resulting in Stacked Bar Charts.
SCMR
All unique values, ( as combined by the XAxis Selector ) would end up on the XAxis
Since there would be single values, Single bars
SCSC
All unique values, ( as combined by the XAxis Selector ) would end up on the XAxis
If there are multiple unique values for each unique value in the column, Stacked Bars otherwise Single Bar
SCMC
All unique values, ( as combined by the XAxis Selector ) would end up on the XAxis
For each of the unique value, there would be a group of bars, one for each of the other column, and each of these bars could further be stacked.
Having a single value per sector, would draw a PIE / Donut, However if you have multiple values across sectors ( i.e. stacked ), you would see a Sunburst.
Mapping
Sectors
Chart Type
MCMR
All columns would be the sectors
If there are multiple unique values, Sunburst else a simple Pie
SCMR
All unique values, ( as combined by the XAxis Selector ) would end up as sectors
Since there would be single values, PIE
SCSC
All unique values, ( as combined by the XAxis Selector ) would end up as sectors
If there are multiple unique values, Sunburst else a simple Pie
SCMC
All unique values, ( as combined by the XAxis Selector ) would end up as sectors
The columns would become the first level Sunburst arc, and their keys would then form the 2nd level arc.
Line Chart
Mapping
Columns
Chart Type
MCMR
All columns would be on the XAxis
Since each column, would have a bunch of keys and values, so Multi Line
SCMR
All unique values, ( as combined by the XAxis Selector ) would end up on the XAxis
Since there would be single values, Single Line
SCSC
All unique values, ( as combined by the XAxis Selector ) would end up on the XAxis
If there are multiple unique values for each unique value in the column, Multi Line else a Single Line
SCMC
All unique values, ( as combined by the XAxis Selector ) would end up on the XAxis
By default the isSummed option would be selected, and works pretty much like MCMR, Otherwise it would get messy.
I am the developer of this Product.
For any suggestions, enhancements, collaboration, you can DM me on
