1:1 Tableau Interview Session | Tableau Training | Tableau QnA | IvyProSchool

[Music] hi google how are you hi man good afternoon i'm fine good afternoon my name is ishani and today we will be talking a little bit about your understanding in tableau but before that can you give me a brief introduction of yourself google i am from tamil nadu i have completed petrol imaginary and i have done my mba in hr and marketing i have worked for a while in a real-time drilling operations company in coaching and there i used to work in a real-time analytics like monitoring the data of drilling and circulation or something like that so i used to provide service to the companies like chevron slumber halliburton then right and due to an injury i should i got took a break and now i pursued a data science and i'm looking for job all right so what made you pursue data science like uh the data was very fascinating while drilling this one it was actually fascinating because uh that we created a small project like autumn custom made uh automatic monitoring process like a project it worked but uh i can't give ideas but at the time i couldn't figure out what was actually happening behind the process i gave an idea but i'm not sure how to make it work so i thought at the time it i found it also also interesting so i googled about it and came to know the data science was behind these and many other things so i thought i would take a trip into my head very good good good so this uh interview that we are going to have is uh from it's it is again data that we're talking about but not predictions we're going to talk about the visualization portion of it sure right and as you're aware tableau is one of the market leaders uh in this particular industry all right um so let's start with the first question if you could explain me that what type of joints are used in blending there's a concept of blending in tableau so what type of joints do blending use in bending the default is left channel default is left join absolutely correct all right and uh do you understand what is hierarchy yes on hierarchy we used to categorize like we used to put in uh group them and into a single club we used to club the dimensions into a single thing in a hierarchical order all right and why do you think we would do something like that like what is the purpose of uh putting the different dimensions into one hierarchy we may have some related dimensions like uh can i say something with the examples now sure like uh some sometimes we used to have a region state and a city at the time we can club them into a single thing with the use of hierarchy and we can use to display it visually more appealing and easily related to the audience plus i think drilling down the data becomes a little easier when you have a hierarchy drilling becomes easier okay uh have you worked in the analytics tab in tableau like there is a data type and there's an analytics tab yes i have worked on it to find the correlations and to find the clubbing process on the grouping process and okay so for analysis i walk down that part so since you're talking about clubbing and grouping and you're talking about correlation so can you explain me what clustering is all about uh clustering is grouping the data of a similar kind into a particular cluster like whether some example means we used to correlate some data with two meshes we used to correlate some data and to find relation at the time we can cluster those datas and find the similarities between them so it is used to find the similarities and group them to find the similar group of similar data okay right so is it like a table you decide how many clusters it can make or even you can tell how many clusters you're looking for uh the time of creation itself which shows how many clusters will be created like uh it is uh it takes create automatic mm it creates automatically all right um let's say if you have uh multiple data sets right but maybe in the same workbook or it may be same different sources of data and we want to work on those data sets and create visualization like we want to put them all the data together and do the visualization so what kind of techniques can we use to merge these data sets um what is the actual purpose for doing this one because based on that uh we can do blending for from different workbooks all right we can use join joins correct and whatever and union union perfect perfect so you can combine the data sources depending on the need like you rightly said yes it would depend on the name that we will do a join or a union so can you explain me what is the difference between a join and a union um join is usually i have it's automatically happened previous in previous tableau version we have to make a specific custom made design to get it into a specific joint but now tableau has evolved and they made it easy to the users to automatically get the left join by default and we can also change the join but in union we are placing uh we are concatenating or something like that we are stacking the data sets uh one at the top and one at the bottom like okay um can we say can we talk in terms of how we are merging in terms of rows and columns like is join merging the data on a row level or a column level and is union what is union difference between union and join when it comes next in a single sheet it adds the two tables uh into a single table in join it is actually happening behind the process but we can uh use it in the sheets to make the view visible okay all right um have you heard about or have you used sets yes can you tell me what are sets why they based on the um based on certain conditions we used to filter the state it is used like uh for filtering process or something like that we can use it for like uh we can cl uh group the data like based on certain conditions like top 10 customers uh who are uh meeting the condition of sales above 10 000 or 10 000 or something like that we can use to create sets and we can distinguish them with the use of multiple cells we can also create a combined sets in that we can perform different type of operations okay so let me elaborate a little bit more on this thing so when you talk about sets uh you said it helps us to group the data to analyze the data right what are we grouping are we grouping based on dimension we are grouping based on measures what are we doing dimension dimension so no measures are involved when we create sets yes it's involved we are using the measures as values but we are actually creating the sets the use of dimensions we are creating a sets on dimensions but we're using measures yes can i have more than one measures based on which i can create a set i can yes one all right and what about parameters parameters is a you know a very good feature of tableau i personally really like parameters so what are parameters parameters are usually gives the user to visualize the data more easily and their preferrable way like we can create a parameters to sort the data we can have a multiple measures in the same data in the same view and we can with the help of parameters we can sort it out uh sort by quantity or by price or something else of whichever measure which we are used in the view it is similar to kind of filter map it's a filtering process it can be used for filtering process mainly parameters all right um let's say if i wanted to analyze um age okay if there's a measure called age if there's a field called age and i would like to run my analysis on age uh what would you do in tableau to you know do a useful analysis to this measure i would use uh build smart bins is the best option to make the measure measure values two dimensions because we can divide we can create slabs like uh from zero to ten uh particular age and ten to twenty likewise we can create different slabs and we can get the values and views with the help of bins with the help of bins can i bench dimension yes can i do binning on dimensions or is it only for measures that i can do billing it's mainly used in measures i'm not sure whether it can be it can be used in dimensions but i haven't used much no not used in dimensions it's only for measures okay okay and if i want to uh you know not exactly binning billing is basically in numbers you're making intervals but if i want to group the items of a dimension like for example let's take um i have uh you know operations in uh many part of india and where there are various cities which are there and i want to you know group certain cities and let's say uh call it region one i want to group another set of cities call it region two so what can i do in that case like if i want if this is a scenario i want i want my visualization i've been given city data city wise data i don't have anything called region defined in my data set and i want to uh show it on my visualization as a region where region one have would have predefined number of necessities like say delhi is there and chandigarh is there things like that so in what in which feature of tableau would you use you can use it by grouping them uh correct you would use grouping over there all right uh so there are various types of you know when you're working in the sheet portion of tableau they are filtered area there is something called a marked card in max car there is a specific card called color card right so can you explain me why or what is the use of a color card and the color card is basically used to differentiate the categories within a field because a field can can have a multiple categories like can i say with some examples like in the sales of our statement items we can get uh like a books uh paper books papers and pens and some other type there are multiple categories we can show the viewers our users with the differentiation like with the use of colors we can differentiate how much sales has happened with uh stationery and with the books and with the in the year graph like that we can show them we can it's a basically used to differentiate it used to it is used for differentiation all right do you think colors are important or uh you should make it visually easy to understand and interpretable so it is not uh it has main importance in the visualization part only visually it becomes appealing and easy to understand all right okay okay so i wanted to ask you that if this is the data set that i have in which i have got the region uh the state and then i've got a main literacy rate and a female literacy rate i want to create a view where i can pick one state let's say i pick delhi and in delhi i want to get a pie chart which shows me what is the percentage of male and female in terms of literacy how would you proceed on that in tableau we can't use this kind of uh data to visually present those uh the thing you asked them because uh uh in tableau usually use to like gender male and female if we have it in a single column we can use we can use it in w to present for that we have to in the data source we have to merge those two columns two columns into a single column and how do we what tool of pivot word will you use for it um i'm not sure about the exact uh word but just tell me the process just tell me the process if you don't remember the terminology that's fine so what is the process that you do we used to select those column and we have used convert it to a single column and then we use it in view then you can use in the view right it's called pivot so your process is clear you know what you're doing but you forgot the terminology that's fine that's fine it's called pivoting so you are absolutely correct in a tableau if your uh headers are given which are supposed to be in one column i couldn't i'm not i will not be able to put it in a pie chart so i need to first convert into a single column of male and female and one column of literacy rate then i will be able to create a pie chart out of it good good okay now let's say um so this pivoting is one way of preparing the data right is there any other tool in tableau that you know of which helps me to you know clean up the data prepare the data for my actual analysis there is an uh at tool data source page absolutely it can be used to clean the data to remove the null values and the blank rows and it used to prep the data it is it is automatically it does in seconds so it can be used to clean up the data clean up the data that's good good all right in the beginning when we started i asked you about blending and what kind of you know join is there in blending but can you explain me why because what i understand and what you said that blending and joining is joining at the end of the day enjoins your various type of voice but blending is only left joint but can you explain me when should i blend the data versus join the data like when should i use blending basically when the level of details are not uh equal or it's different when the level of details are different we have to use blending because can you give me an example of the two tables where we can explain the level of detail because in the truth a table may contain a category and the other table may not contain a category but it contains a subcategory at that time the level of detailing is different but they are relatable so at the time we can use blending uh to visually present the data okay so what is the process of what does blending actually do in like internally and tableau what it is doing it actually it writes sql code behind the process behind the process and it converts into a table into a specific format and both tables are joined good good good all right okay let's talk about uh filtering so do you know what is the context filters what are context filters when we use when we use multiple filters tablet doesn't work so at the time we have to use add to context to make the second second or third filter to work okay so for example let's say let's take a scenario i've got a data for countries and within the countries i have a data for states right so i have data for country states and there's sales right so if i want to create a view where i want to find the top five states for each country and i give the control to the user through the filters to pick one country at a time so let's say if i pick usa i should get the top five states for usa in terms of sales so can you tell me the process at first we have to create a parameter let's say we are creating filters filters we can put kind of increasing filters and uh the person is picking filters you know country from filter in the filters we can create by top and there is a tab like a top parameters top tab in the tab we can mention how many number of uh fields we have want to display so in that thing we can make it as five and we can select the city and uh the thing and then we can make it the user make it visible okay so now like in this scenario we have used two filters correct one i've used in country and one you said we will put it on the state and we'll use the tab of top and put the top five over there right now there are two filters on which field will you put the context filter like which one will be add to context the other one then the top i wish we couldn't we should not use on the on the top one uh other one so there are two filters okay there is a country filter and there is a state filter filter in the country in the country printer okay good um you also spoke about a while back when we were talking about clustering you were talking about correlations right so what kind of charts do you typically use for correlations and to create a scatter chart how many measures minimum measures do we require minimum two measures and one dimensions okay all right and um let's say let's say if you have uh you know data for different months let's say i have sales data and each sales data is recorded on different months like january has one sheet feb has one sheet much as one sheet like this i have got generally to december all the sheets with the data now i want to do a analysis of my you know whatever measures that are there i want to do analysis on a yearly basis okay but my data is segregated into 12 different sheets so how will i do my yearly analysis um i think the union would be the better option for this process now i haven't tried this exact scenario but i think after with my knowledge and experience i think the union would be right you have to combine all the data set one on top of other so get it in one table and then we can do it correct correct okay um let's say if i have uh in one of the columns i have the customer's name okay one of my columns is customer name where i have the customer first name i've got its middle name and i've got the last name of the customer now my intention is that i would want to find out the first name of the customers because let's say i'm in a business where we do customize gifting so you know few of the gifts i want to already pre-create to the typical name of the people who come and buy at the store okay so for that purpose what i need to know is the uh occurrence of each first name the number of each first name right so i'll repeat my question my question is that i have a column in which i've got the customer's first name middle name last name in one single column my intention is that i want to find out let's say a google okay how many googles am i you know buying from me i want to understand what the what are the common names of my customers so that i can pre-create some grips and keep it so how would you do it uh for the thing we have to segregate uh separate the first name from the other name we have to use a custom split and the data source sheet okay data source we can specify the delimiter where we have to separate and how much in the first attempt or the second atom we can choose like otherwise if we choose one first attempt and a delimiter as if they are having a space between their names space we can separate into a single column we can get only the single first name okay so you've done one step of it that you've got the first names out of the entire name my thing was first name now what how will you do the second part to it where i want to understand how many of each are there and then we can convert this field into we can use it as a dimension and we can use the count option to aggregate function to find the number of uses correct okay um tell me about table calculations have you heard about table calculations yes one i use the most in the analytical parts okay similar to that it's similar to show as value in excel because we can get the all the similar features like uh percentage for total percentage and the calculations like uh subtotals and something like that we can get a series of calculation parts using so what is the difference between double table calculation and a calculated feed like how would i differentiate when should i use a calculated field there's a calculated field also in tableau so what is the difference between the two calculated field is created by ourselves based on a certain condition one okay calculated but uh the table calculation is already is performed on the dimensions and the measures which we use to create a value like we have like for an example in a view we are creating a company's name and the excels that we can we want to show the user or our manager that what is the distribution of growth or sales in the particular uh company at the time we can convert into it into a total of percentage total percentage is of sales calculation oh yes one this is what they will calculate okay now let me give you another scenario let's say um we have various products that are manufactured in the company and we have recorded year by sales of it okay so i have products let's say 200 products and i've got five years data i've got the sales value of those products now i want to understand how much growth has been there for each product over a period of time how would i find that out we can use the line graph to i want the values i'm more interested in what was the change for example if it was like say 200 000 in uh year 2017 in 2018 it was let's say 250 000 so what was the change or what was the increase uh we can use it by if we can create a filter also known by for shifting from particular uh year to year or by we can create a parameters like from start date to end date we can uh divide those status and i'll repeat i'll repeat what i'm looking for is um imagine a matrix okay okay one column the rows that you have consists of the product names okay right in the columns i've got the ears okay right 17 18 19 20. now i want to understand from 17 to 18 what was the percentage increase or decrease in the sales value of that product and we can select an option from previous uh thing we can subtract correct so that's the later on your growth correct we can find the difference so basically in um it is going to be in table calculation year on year growth yes all right uh what kind of projects have you done google on tableau can you give me a few tablets yes maintenance case study and i have worked on some other minor projects studies and projects okay that's great uh any challenges that you faced while working on uh the data sets like any challenges in tableau that you faced i'm not sure whether i can say this or not i got too comfortable with the tablet tool and it was uh automatically i got a flow with that so i didn't face any much trouble with the tablet part and while creating a visually appealing on the visually attractive display was the major trouble i got i think so even though i created a view with a fascinating color and fonts or something like that i couldn't get the full satisfaction from those views so i'm still still trying to get a classic view something like that all right all right great one last question google that is let's say if i uh hire you as a consultant for my company okay and uh i give you my data set and say okay google you don't have you okay why can you create a dashboard for me so how would you proceed on doing this work but first uh i'll i'll ask you some questions like uh for what purpose we are doing this what we are trying to achieve and what are the feels and what they depressing and what are the objectives we are trying to get like that i used to get some data from you and then with those status i used to create a objectives and based on those objectives i try to perform the visualization part very nice very nice good good good all right thank you google thank you for your time it was a pleasure talking to you thank you take care bye [Music]

Channel: IvyProSchool

Views: 48,641

Keywords: analytics, datascience

Id: i6zhYhDN5CQ

Length: 27min 55sec (1675 seconds)

Published: Tue Dec 08 2020