Week 1 - Is email address Structured or unstructured data ?

[Darria Osline]

Email address contain text,numbers and special characters.How does that come in structured data.Kindly explain.

[Anand Iyer]

characters uses don't define structured and unstructured data.  The pattern does.  If there's a clearly identifiable, and usable pattern in data, it's structured.

Emails are short (can't exceed 320 characters) and always have a part that goes before (user part) and after (domain part) )@ symbol.  THat's a pattern good enough for our purpose.

[Antony]

Lets say you have a dataset 1 in the below format

CandidateName | Contactnum | Pincode | Email | Address | City

Darria | 1234567890 | 5860027 | Darria @something.com | #hno 1/22, Abcd road | some city 

Dataset 1 is structured, since you are able to identify every attribute of the observation provided. 

Lets say you have a dataset 2 in the below format

Candidate Information

Darria who is from some city , residing at #hno 1/22, Abcd road , can be reached at 1234567890 mobile and email id Darria @something.com

Dataset 2 is unstructured, because you cannot do any analysis on top of this text field. You will have to first make it structured , similar to dataset 1.

Whether dataset is structured or not , is decided on the context of what analysis you want to perform and whether it can be achieved from the 'dataset in the raw form' that was provided to you.

[Darria Osline]

Thank you guys.

[Anand Iyer]

Very good explanation, Antony!

On Thu, Nov 12, 2020 at 11:27 AM Antony <royd...@gmail.com> wrote:

--
You received this message because you are subscribed to a topic in the Google Groups "Discussion forum for Statistics for Data Science I" group.
To unsubscribe from this topic, visit https://groups.google.com/a/nptel.iitm.ac.in/d/topic/ma1002-discuss/KpMMg0AcXXk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ma1002-discus...@nptel.iitm.ac.in.
To view this discussion on the web visit https://groups.google.com/a/nptel.iitm.ac.in/d/msgid/ma1002-discuss/5230ddb8-cfc2-49f0-947e-add3136cdfd4o%40nptel.iitm.ac.in.

--

Cheers,

Anand

[Antony]

@Anand, its my daily bread :P

Just to add on to my explanation , lets say you have dataset 3

Candidate | Email Id

Darria | Darria123 at gmail . com

Anand | Anand456 at yahoo . com

The analysis you want to perform , requires you to find the email provider (gmail , yahoo etc)  of the candidate .  This is not straight away possible on the email field , but since you know the general format of an email id ( something @ something . something ) . YOu can do basic string operation on email field to find the email provider. So this makes "email id", in this dataset a semi structured data field. 

Examples of structured files : files where data is in rows/columns format 

Examples of unstructured files : log (text) files, audio/video files

Examples of unstructured data : xml files, email ids etc

On Tuesday, November 10, 2020 at 10:39:00 PM UTC+5:30, Darria Osline wrote:

[Anand Iyer]

ah, you brought a new variation now called semi-structured.

given the options structured or unstructured, you'll classify email address as structured, right? Or do you have a different opinion?

BTW, are you working as a data analyst/scientist, @Antony?

--

You received this message because you are subscribed to a topic in the Google Groups "Discussion forum for Statistics for Data Science I" group.
To unsubscribe from this topic, visit https://groups.google.com/a/nptel.iitm.ac.in/d/topic/ma1002-discuss/KpMMg0AcXXk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ma1002-discus...@nptel.iitm.ac.in.

To view this discussion on the web visit https://groups.google.com/a/nptel.iitm.ac.in/d/msgid/ma1002-discuss/38fc23df-ee4a-471d-aaf3-c72d93eb691co%40nptel.iitm.ac.in.

--

Cheers,

Anand

[Antony]

As per sample dataset 3, if I have to do some analysis on email provider and semi structured  is not one of the available options , I will choose structured.  

Even with semi structured given as one of the option, I would think this is one of those questions which will lead to ambiguity depending on how you personally interpret it. Because doing a basic string operation is no big deal in the given example. But if the file was in xml format, it definitely has to be semi structured. 

@anand I do work in data analytics field, but mostly into descriptive analytics. This course is me trying to get into more advanced analytics. But if mock test is anything to go by , I will most probably say tata byebye in qualifier since I need my excel to do math , I cant work pen and paper anymore even for single digit calculations :)) . Not to mention aspirations and labiowhat in english.  

On Tuesday, November 10, 2020 at 10:39:00 PM UTC+5:30, Darria Osline wrote: