BLOGS

ByHariharan Rajendran

Linear Regression

Before starting with linear regression, let us try to understand the machine learning concepts.

As of today, machine learning is one of the trending technologies as the applications are started to consume real-time data and deliver machine learning services.

As part of machine learning, we have three different types of machine learning models available.

  1. Supervised Learning
  2. Unsupervised Learning
  3. Reinforcement learning

In Supervised Learning, the system tries to learn from the previous examples that are given.

Under supervised learning, we have classification and regression models.

One of the common and simple techniques for predicting a continuous variable is called linear regression.

It assumes a linear relationship between the outcome and the predictor variables.

The linear regression equation can be written as y = b0 + b*x + e, where:

  1. b0 is the intercept,
  2. b is the regression weight or coefficient associated with the predictor variable x.
  3. e is the residual error

This post covers the linear regression on excel and R studio.

Linear Regression on Excel

Follow the below steps to perform the linear regression in excel

Step 1: Consider you have year wise sales amount. You are in the situation to predict the sales amount for the year 2011.

Step 2: Create a scatter plot like below with Year and Sales Amount

Step 3: Right-click on the line and choose trendline.

Step 4: Display the equation on the trendline by selecting the checkbox on the Format trendline properties.

Step 5: Predict the value using the generated formula for existing values.

Step 6: Predict for the year 2011.

As you can see the values are not 100% correct, we need to adjust the con-efficient and reduce the r2 values to get the perfect result.

ByHariharan Rajendran

Power BI Report using Shared Dataset [Detailed Step by Step]

In my previous post, I have explained the possibility of using SSRS shared dataset as a source for Power BI report. Refer my previous post here.

In this post, I am giving you the detailed step by step procedure to work with the shared dataset in Power BI.

Follow the below steps,

Step 1: Open Power BI Report Server portal URL. For example, the URL is like below.

http://localhost/Reports

Step 2: Add the following extension with the above URL. /api/v2.0/datasets. It will list out all the datasets with ID as like below.

http://localhost/Reports/api/v2.0/datasets

Step 3: Filter the dataset with ID. Here I used the second dataset id.

http://localhost/Reports/api/v2.0/datasets(datasetid)/data

My URL, http://lab/ReportsPBI/api/v2.0/datasets(ca502b81-c88d-4646-af32-ca08d9b4a1f4)/data

Step 4: Open Power BI Desktop and choose OData feed data connector

Step 5: Pass the above URL as like below.

Step 6: Choose advance and select “Include open type columns” checkbox to get all the columns of the dataset.

Step 7: Click Ok and select Edit Option on next screen.

Step 8: Select the icon near the “More Columns” as like below. It will list out all the columns.

Step 9: Uncheck the “Use original column name as a prefix” and click ok. Finally click “Close & Apply”.

Step 10: Choose the fields and create a report.

Thanks. Let me know if you face any issues.

ByDr. SubraMANI Paramasivam

Empowering Every Person in the planet, with “Awareness Enabled Reports”

Following Satya’s quote “Empowering Every Person in the Planet”, in 2016 Microsoft Inspire conference, have certainly left my mind with unstoppable beats. I certainly understood what Satya meant, but I still thought about doing something like this, with my own upgraded version of Data Awareness Programme (which I started in 2014). But at that time, all I had was Power View, Power Map, Power Pivot as Excel add-ons and tried my best, to do the awareness programmes in remote villages, by mingling with the villagers and collecting their own data and showing some visuals back to them. I did this to improve their life style, find more time for personal and have better earnings. Though the main target was students, I hoped this message would spread to their friends, family and others in the remote places.

Following the release of full version of Power BI, I now have a fully working site, with living “Awareness Enabled Reports”, from sleeping open data sources (taken from various Gov/Non-Gov sites).

I managed to get this far, with a simple equation of A + B + C + D = E (EmpoweringEveryPerson.com (EEP)). Let me explain this in detail.

Following this, I managed to extract data from various open data sources and identified the most global challenges, that we are encountering and/or going to be a major threat in near future.

With our all time favourite reporting tool, “Microsoft Power BI“, published some “Awareness Enabled Reports” to www.EmpoweringEveryPerson.com site and categorized them with regional, national and global challenges, for easy manoeuvring within the EEP site.

 

This EEP site currently has 3 simple goals.

  1. View the “Awareness Enabled Reports” worldwide in any devices, by categorizing Regional or Global challenges.
  2. Submitting another “Awareness Enabled Report” as a Developer with some guidance. Also listed, Global Challenge Topics to select from open data sources and provided some tips to convince the selection committee and finally submitting the story.

  1. Promote EEP
    1. Provided options to Promote EEP as  a Developer, User Group Leader & End user.

Below screenshot shows categorization of reports by UK.

 

PROMOTE & PARTICIPATE

As per above equation, ‘D’ is the support that we need from you, to promote in any of the following ways.

1. DEVELOPERS

Develop “Awareness Enabled Reports” for all listed Global Challenge topics and submit your data story to EEP site and once published, tweet / share / post in social media and spread the awareness.

2. COMMUNITY LEADERS

A request to all Community group leaders, to spend at least a minute by starting or ending your user / local / online group sessions by introducing / re-introducing, this website and showcasing the opportunity to all attendees, to build and submit their own data story with “Awareness Enabled Reports“.

3. END USERS / VOLUNTEERS

Every time you see a new “Awareness Enabled Report“, do tweet / share / post in social media and support to spread the awareness.

 

Thanks in advance for your support and thanks for your time reading through this far, to create awareness with “Awareness Enabled Reports“.

ByHariharan Rajendran

How to handle date in R

Date data type is very important for all programming languages. To handle the date in a proper way, we need to apply some formatting logic.

This is the case for R programming language as well. This post explains you, how we can handle the Date data in R. There are different functions available in R to handle date.

To get the today’s date, use Sys.Date()

Sys.Date()

> Sys.Date()

[1] “2018-05-22”

 

To get date and time, use date()

> date()[1] “Tue May 22 15:29:18 2018”

 

Check the below list of symbols to play with date and time format.

Symbol Meaning Example
%d day as a number (0-31) 01-31
%a
%A
abbreviated weekday
unabbreviated weekday
Mon
Monday
%m month (00-12) 00-12
%b
%B
abbreviated month
unabbreviated month
Jan
January
%y
%Y
2-digit year
4-digit year
07
2007

 

> today <- Sys.Date()> format(today, format=”%B %d %Y”)[1] “May 22 2018”

 

Date Conversion

In a real-time scenario, we will not get date as date data type. Always, we need to convert to proper date data type. To convert to date data type, we need to use as.Date() function.

Syntax

as.Date(x, “format”)

x – Field or column

format – use the above symbols to frame the proper format.

 

DateVec<- c(“01/05/2018”, “08/16/2018”)

DateVec

[1] “01/05/2018” “08/16/2018”

summary(DateVec)

Length     Class        Mode

2            character   character

dates <- as.Date(DateVec, “%m/%d/%Y”)

dates

[1] “2018-01-05” “2018-08-16”

summary(dates)

Min.      1st Qu.       Median         Mean      3rd Qu.         Max.

“2018-01-05” “2018-03-01” “2018-04-26” “2018-04-26” “2018-06-21” “2018-08-16”

ByDr. SubraMANI Paramasivam

Run Your Python Script

Use the below console to run your python scripts

# Assign value to the variable # a = 5 #Print
Print
ByHariharan Rajendran

R Package dplyr Function vs T-SQL – Part 2

Following my previous post, we have another few functions on dplyr package which will be covered in this post.

Arrange Function

As part of data modelling, we need to sort the data to analyze further.

In T-SQL we can easily perform the sort on the data as like below. It uses “Order by” clause to sort the data.

In R, we need to use arrange function.

#Arrange

DF_Arrange<- DF_select %>%

group_by(State) %>%

summarise(Total = sum(Price)) %>%

arrange(desc(Total))

DF_Arrange

Mutate Function

The next, very common task is to build the calculated column to satisfy the business logic. This can be easily done with the help of “mutate” function.

#mutate

DF_mutate <- DF_select %>%

group_by(State) %>%

summarise(Total = sum(Price)) %>%

mutate(“10%of Total” = Total/10) %>%

arrange(desc(Total))

DF_mutate

ByHariharan Rajendran

R Package dplyr Function vs T-SQL

R is one of the very famous tools to handle the data science projects because it has all the capabilities right from the extracting the data from different sources, data modelling and transformation, data visualization and finally building machine learning models using the data.

This post explains how the data modelling can be done with R using “dplyr” package.

To make it easy, let me compare this dplyr function with T-SQL.

First, let us load the data into R studio.

fulldata<-read.csv(“D:\\Projects 2018\\DSKA\\DataWranglingDemo.csv”)

fulldata

Select Function

When we have a large dataset with more than 1000 columns, if we need only certain columns then we can use “SELECT” option to choose the specific columns. Check the below example,

 

The same option is available in R with dplyr package.

#install.packages(“dplyr”)

require(“dplyr”)

#select

DF_select <- fulldata %>% select (State,Price)

DF_select

Filter Function

In T-SQL, we have where condition to filter any data with different conditions.

The same filter option is available with dplyr functions.

#filter

DF_select %>% filter(State==”Alabama”)

Same as T-SQL, we have group by, summarise functions are available in R

Group by Function

#Group by

DF_Group <- DF_select %>% group_by(State)

DF_Group

Summarise Function

#Summarise

DF_Sum<- DF_select %>% group_by(State) %>%

summarise(Total = sum(Price))

DF_Sum