یادگیری تحلیل داده از صفر با AI | بدون این ۶ قدم، Data Analyst نمیشی! — Transcript
Full transcript
- 0:01Why, now that AI can write SQL, do you
- 0:04see companies like Digikala and Snapp
- 0:06looking for this very skill when you
- 0:08open data analyst job postings?
- 0:16Well, why do companies with full access
- 0:18to AI still look for people who can
- 0:20write queries and know SQL?
- 0:24In these same ads, what else are they
- 0:26asking for alongside SQL that AI lacks,
- 0:28and where should you, with no experience
- 0:31, start so that one of these job ads
- 0:33actually responds to you?
- 0:35By the end of this video, you will have
- 0:37a complete roadmap; from the very first
- 0:39skill to your first interview. And it's
- 0:42all based on what companies in Iran have
- 0:44actually written in their job
- 0:45descriptions. Even if you haven't
- 0:47written a single line of code until now
- 0:49, it doesn't matter at all; because we
- 0:50are going to go forward step by step
- 0:52together.
- 0:53For those seeing my video for the first
- 0:55time, I’m, and I teach you
- 0:57everything from the world of AI that
- 0:59impacts my own life in simple terms.
- 1:02Well, this path we are going to take
- 1:04together has 6 steps. Step one: Excel,
- 1:07getting familiar with data without even
- 1:09a single line of code. Step two: SQL,
- 1:12the most important skill for this job.
- 1:15Step three: Basic statistics, so you
- 1:17understand if a number is truly
- 1:19meaningful or not.
- 1:21Step four: Dashboards or Power BI.
- 1:25Step five: Python for when data gets big
- 1:28.
- 1:30Step six: AI, which makes you faster
- 1:32rather than replacing you.
- 1:35After that, I'll tell you about a skill
- 1:37more important than all of these, and
- 1:39finally, project building and finding a
- 1:41job. For every step, I will tell you
- 1:43three things. First, why is it necessary
- 1:46?
- 1:47Second, exactly what to learn and in
- 1:49what order.
- 1:51Third, when you are ready for the next
- 1:53step.
- 1:54Let me tell you one point before
- 1:55starting.
- 1:56From day one, practice on real data.
- 1:59Not fake examples.
- 2:01There is a site called Kaggle that is
- 2:03full of real, free data.
- 2:06It has data from a Brazilian online
- 2:08store with about 100,000 real orders;
- 2:10what did they buy? How much did they buy
- 2:13, when did it arrive, and what rating
- 2:15did the customer give?
- 2:17All the exercises in this tutorial are
- 2:19based on this file, and I will put the
- 2:21link in the description of this video
- 2:23for you to use.
- 2:25Step one: Excel. Why Excel first?
- 2:28Because it’s the simplest way to get
- 2:30familiar with data. It doesn’t require
- 2:32code. Almost every company works with it
- 2:34.
- 2:35Digikala has asked for Excel and SQL
- 2:37together for one of its data analyst
- 2:39roles. Exactly. Now, what should you
- 2:41learn? In the exact order I tell you.
- 2:44First, sorting and filtering.
- 2:46Sort and filter.
- 2:48For example, being able to see only the
- 2:50orders from a single month.
- 2:54Next, some basic formulas like sum,
- 2:56average, and conditional logic, plus
- 2:59joining two tables via a common column (
- 3:01the latter is called VLOOKUP); then data
- 3:04cleaning, which means finding and fixing
- 3:07duplicate rows or empty cells; finally,
- 3:09a summary table that aggregates
- 3:12thousands of rows with a few clicks,
- 3:14known as a Pivot Table, along with a
- 3:16chart.
- 3:19Let me mention a point specific to
- 3:21Persian data. When working in an Iranian
- 3:24company, the date is usually Persian (
- 3:26Shamsi). In some places, numbers are
- 3:28typed in Persian, while in others, they
- 3:31are in English.
- 3:34The name of a city, like Tabriz, might
- 3:37be entered with an Arabic 'ye' or a
- 3:39Persian one; the computer considers
- 3:41these as two different cities.
- 3:45So, we must take data cleaning very
- 3:48seriously. If the data is dirty, any
- 3:50analysis you perform afterwards will
- 3:52likely turn out to be wrong. Now, when
- 3:55should you move on to the next step?
- 3:58When you can open the orders file in
- 4:00Excel, use a Pivot Table to show how
- 4:02many orders were recorded each month,
- 4:04and have a chart next to it.
- 4:07Step two: SQL. Why SQL? It is exactly
- 4:10what we saw at the beginning of the
- 4:12video in job ads for Digikala, SnappFood
- 4:15, Snapp, Basalam, Alibaba—it's mentioned
- 4:17almost everywhere. First, let me tell
- 4:19you what it is. A company's data is not
- 4:22in Excel.
- 4:23It is stored in a place where several
- 4:25tables are kept together.
- 4:28A table for orders, for instance, a
- 4:30table for customers, a table for
- 4:32products—like several accounting ledgers
- 4:34placed side by side.
- 4:37This collection is called a database.
- 4:39SQL is the language you use to ask
- 4:41questions of these ledgers. Each
- 4:44question is called a query. Now, what
- 4:46exactly do you need to learn?
- 4:49In this specific order: First, selecting
- 4:52and filtering (SELECT and WHERE), like "
- 4:54give me last month's orders"; then
- 4:56aggregating and grouping (SUM, GROUP BY
- 4:59), for example, "give me the total sales
- 5:01broken down by product category"; then
- 5:04joining tables together, connecting
- 5:06orders to products to see what each
- 5:08order contained; this is called a JOIN,
- 5:11which is a term specifically mentioned
- 5:13in job ads.
- 5:17Finally, once you are comfortable
- 5:19writing these, move on to nested queries
- 5:21and window functions.
- 5:24For practice, sites like LeetCode and
- 5:26Stratascratch are great resources that
- 5:28offer free SQL questions.
- 5:31Now, when should we move on to the next
- 5:33step?
- 5:34When you can answer a question using a
- 5:36single query on that Kaggle data, for
- 5:38example, "which product category had the
- 5:40highest sales?"
- 5:42; this question requires all three
- 5:44skills together. Before we continue, my
- 5:46friends, I’d like to ask that if this
- 5:48tutorial is useful to you, please like
- 5:51the video, share it with your friends,
- 5:53subscribe to my channel, and hit the
- 5:55bell icon so you don't miss future
- 5:57tutorials; because I create tutorials
- 5:59every week that can have a positive
- 6:01impact on your life. Step three: basic
- 6:03statistics. Now, why statistics? Because
- 6:06up to this point you’ve learned to
- 6:08extract numbers, now you need to
- 6:09understand if those numbers are actually
- 6:11meaningful or not.
- 6:12For example, Basalam specifically asks
- 6:15for the cause in the job posting from
- 6:17anyone who wants to fill out the form.
- 6:21You don't need to be a statistician, but
- 6:23you must know these three; these three
- 6:26are enough: first, average or mean, and
- 6:28median.
- 6:30I’ll give an example so you understand:
- 6:32imagine a 10-person company where the
- 6:34manager makes 10 times more than the
- 6:36others. The average salary comes out
- 6:38high, but it doesn't represent the
- 6:40salary of most people.
- 6:42The median is the middle number when you
- 6:44sort everyone from low to high; here,
- 6:46the median is more realistic.
- 6:50Second, dispersion or variability.
- 6:54This means how much the numbers go up
- 6:56and down.
- 6:58If sales naturally fluctuate by 10 each
- 7:00month, a drop of, say, 5 isn't something
- 7:03strange.
- 7:06Third, correlation, which means two
- 7:08things go up and down together.
- 7:12A common mistake happens here; for
- 7:14example, in the summer, ice cream sales
- 7:17go up, and air conditioner sales go up
- 7:19too.
- 7:20These two have a correlation, but ice
- 7:22cream doesn't cause someone to buy an AC
- 7:25; both are because of the hot weather.
- 7:27This is called correlation, not
- 7:29causation (correlation is not causation
- 7:32).
- 7:33Finally, you should also know one more
- 7:35term called hypothesis testing; it means
- 7:38checking if a change is larger than the
- 7:40normal fluctuation or not.
- 7:43You don't need to memorize the formulas
- 7:45at first, you just need to know what it
- 7:47is at all. When should you go to the
- 7:49next step?
- 7:51When you can calculate the average and
- 7:53median of order amounts in that same
- 7:54data and explain why they are different
- 7:56from each other.
- 7:58Dashboard with Power BI; why a dashboard
- 8:01? Because managers don't have time to
- 8:03read tables.
- 8:04A dashboard is a single page that shows
- 8:07important business numbers in one place
- 8:09, live; like the speedometer in a car
- 8:11where speed and fuel are always in front
- 8:13of our eyes. What is the tool for it?
- 8:16In job postings, the name you usually
- 8:19see most often is Power BI; Snapp, Asan
- 8:22Pardakht, and Alibaba all use such tools
- 8:25.
- 8:28Power BI, but Tableau and Metabase do
- 8:30the same thing (you can build dashboards
- 8:32with them), but they are used less.
- 8:36Now, what do you need to learn for this
- 8:38?
- 8:41Follow these steps in this exact order:
- 8:44first install, then import your data
- 8:46file (the desktop version of Power BI is
- 8:49free), create a basic chart—like sales
- 8:51by month or sales by category—and
- 8:53finally, add a filter to the page so the
- 8:56manager can choose what to see; this
- 8:58filter is called a slicer.
- 9:03One important note: you only learn to
- 9:05build a dashboard by doing it, watching
- 9:07videos alone isn't enough. So, when
- 9:09should you move on to the next step?
- 9:11When you have created a one-page
- 9:13dashboard using Kaggle data that shows
- 9:15monthly sales and top-selling categories
- 9:17.
- 9:18Step five: Python. Now, why Python?
- 9:22There are two reasons.
- 9:24First, an Excel sheet can't hold more
- 9:26than about a million rows.
- 9:29When data gets larger, Excel falls short
- 9:31.
- 9:32The more important reason is that even
- 9:34when AI writes Python code for you, you
- 9:36need to be able to read and understand
- 9:38what it’s doing, though you don’t need
- 9:40to be a programmer. So, what exactly do
- 9:43you need to learn? In this order? First
- 9:45, Python basics: variables, lists, and
- 9:48conditions; all of that is enough.
- 9:51Then libraries—Python has a set of
- 9:53pre-made toolkits so you don’t have to
- 9:55sit there and write everything from
- 9:57scratch.
- 9:59These are called libraries.
- 10:01For data analysis, two of them are
- 10:03enough for you. Pandas is for working
- 10:05with tables, and NumPy is for working
- 10:07with numbers. Companies like Digikala
- 10:10and Digishahr mentioned these two
- 10:11libraries by name in their job postings
- 10:14. Finally, where should you write your
- 10:16code?
- 10:17There’s a notebook where you can write
- 10:18your code in pieces and see the result
- 10:20of each piece right there.
- 10:24These notebooks are called Jupyter
- 10:26Notebooks, and you can install them on
- 10:28your own system for free without needing
- 10:30a powerful computer.
- 10:35The online version is Google Colab,
- 10:37which we used, and it is currently
- 10:39accessible from Iran.
- 10:42Now, when should we move on to the next
- 10:44step?
- 10:45When you can recreate the pivot table
- 10:47from the first step—the number of orders
- 10:50each month—using just a few lines of
- 10:52Pandas. Step six: Artificial
- 10:54Intelligence. Now for the final step and
- 10:57the answer to the video's first question
- 10:59: don't just see AI as a line on your
- 11:01resume; it should become your daily
- 11:03assistant.
- 11:04So, what is it good for?
- 11:08Learning every stage of this data
- 11:10analysis path we are talking about.
- 11:14Writing the first version of a query.
- 11:16You give it your question in Persian,
- 11:18and it delivers a query; explaining
- 11:20errors—for example, if your Python code
- 11:23has an error, you give it the error and
- 11:25ask what it means. Where is the problem
- 11:27? Repetitive tasks!
- 11:30If you download a report every week,
- 11:32repeat the same cleaning, and keep
- 11:34sending it to your manager, ask which
- 11:36parts can be handed over to AI to build
- 11:38an automation.
- 11:42All these tasks we’ve been discussing
- 11:44are called automation.
- 11:46But one skill here is more important
- 11:48than all: knowing whether the answer the
- 11:51AI gave is actually correct or not.
- 11:54Let me give you a figure; there’s a
- 11:56famous test called "Bird" where they
- 11:58give AI questions to write SQL correctly
- 12:00.
- 12:02The best models, when working alone, got
- 12:05about 80 percent right, but human
- 12:07experts got about 93 percent right.
- 12:11Meaning, in that test, almost one out of
- 12:14every five queries the AI wrote was
- 12:16incorrect. Let me tell you a common
- 12:18mistake that happens with this very
- 12:20Kaggle data. You ask the AI, for example
- 12:22, to calculate total sales.
- 12:26It joins the payment table to the order
- 12:28items table.
- 12:32Now, if an order has 3 items, the amount
- 12:35for that order is counted three times,
- 12:37making sales higher than reality, and
- 12:39the scary part is the query runs without
- 12:42any errors.
- 12:45The only person who can catch this error
- 12:47is someone who knows SQL themselves.
- 12:50So, the answer to the first question of
- 12:52our video is exactly this. Do companies
- 12:54still want SQL? They want someone who
- 12:57knows SQL because someone needs to check
- 12:59the AI's answer to see if what it wrote
- 13:01is correct or not.
- 13:02Now, when do you learn this step?
- 13:05When you give the second-step question
- 13:07to the AI, compare its query with your
- 13:09own, and if they differ, understand
- 13:11which one is correct and why.
- 13:14The most important skill of all. Well,
- 13:16our 6 steps are finished.
- 13:19But there is a skill you must practice
- 13:21alongside all these steps, and AI will
- 13:23help you the least with this one.
- 13:27That is reaching a decision from numbers
- 13:29.
- 13:31SnappFood wrote in its delivery team ad
- 13:33that your job is to be able to find the
- 13:35root cause of major problems.
- 13:38Sedat also wrote that you should be able
- 13:40to explain complex ideas simply. These
- 13:42points have two parts.
- 13:46The first part, understanding the
- 13:48business itself; AI can say a number has
- 13:50dropped, but it doesn't know what
- 13:52happened outside the table, within the
- 13:54company, to cause that drop.
- 13:58To start, choose one sector and learn
- 14:01its key metrics.
- 14:05For example, if you like online retail,
- 14:07see how many out of every 100 people who
- 14:09open the site actually make a purchase;
- 14:12this number is called the conversion
- 14:14rate.
- 14:17The second part is storytelling.
- 14:20You shouldn't take the result of your
- 14:22analysis to your manager with 20 charts
- 14:24.
- 14:24Take 3 sentences.
- 14:26Let me give you an example using that
- 14:28same Kaggle data.
- 14:30Suppose in your analysis you saw that
- 14:32orders that arrived late received lower
- 14:34scores. You go to your manager and say,
- 14:36here is what I found. Orders that arrive
- 14:39late have lower customer ratings. Why is
- 14:41it important? Dissatisfied customers
- 14:43won't buy from us again. What is my
- 14:45recommendation?
- 14:48I suggest we start by solving the issues
- 14:50in cities with the most delays; use a
- 14:52chart, not 20—we found why it's
- 14:54important and what my proposal is.
- 14:59Now, the projects; you don't need 10,
- 15:01just 3 strong ones, and the good thing
- 15:04is if you've followed the steps with
- 15:06that Kaggle data I mentioned, you're
- 15:08already halfway there.
- 15:10Project one: A sales dashboard with
- 15:12Power BI. Project two: An analysis with
- 15:15SQL, for example, the best-selling
- 15:18categories and seasons. Project three:
- 15:21An analysis with Python that includes a
- 15:23suggestion at the end, like the
- 15:25relationship between delays and customer
- 15:27scores for the next step, or choose
- 15:29based on job ads.
- 15:31For instance, Snapp Market mentions in
- 15:34their ad that experience in e-commerce,
- 15:36fintech, or logistics is a plus.
- 15:39So, choose your projects in one of these
- 15:42areas.
- 15:43And let me tell you one more tip.
- 15:45Don't just dump your code on GitHub and
- 15:47think you're done.
- 15:49For each project, write down what the
- 15:51problem was, what you discovered, and
- 15:53what you suggest.
- 15:55Because employers want to see how you
- 15:57think, not just that you know how to
- 15:58code.
- 15:59Now, let's go to the last step. Finding
- 16:02a job.
- 16:03Don't wait until you've learned
- 16:04everything perfectly.
- 16:06Start sending out resumes as soon as you
- 16:08finish your second project. Don't
- 16:10underestimate internship ads either. For
- 16:12example, DigiShahr wrote for a data
- 16:14analyst intern that experience doesn't
- 16:16matter at all. Basalam also wrote that
- 16:18they are looking for everyone from
- 16:19interns to seniors. For the interview,
- 16:22practice these items.
- 16:25Have a practice and review of SQL and
- 16:27Excel questions, as they are usually the
- 16:29first filters.
- 16:32A business problem that you must solve
- 16:34right there in the interview; this is
- 16:36called a case study, which you should
- 16:38practice with AI.
- 16:41And behavioral questions about teamwork
- 16:43are usually asked.
- 16:48And do one more thing: ask AI to play
- 16:50the interviewer for you; tell it: "You
- 16:52are a data analyst interviewer, give me
- 16:55a case study, and critique and correct
- 16:57my answers."
- 17:00Let's summarize what you need to do.
- 17:03Okay, I'll summarize the whole path for
- 17:05you once. Excel is for getting familiar
- 17:08with data.
- 17:09SQL is a data analyst's most important
- 17:12skill.
- 17:13Basic statistics so you understand if a
- 17:15number is significant or not. Power BI
- 17:18for dashboards.
- 17:20Python for big data and AI as an
- 17:22assistant (provided you check its
- 17:24answers yourself), and alongside all of
- 17:26them is getting from numbers to
- 17:28decisions.
- 17:30If I had to tell you just one thing to
- 17:33start this week: download that Kaggle
- 17:35data, open it in Excel, and build your
- 17:37first pivot table.
- 17:40AI is changing the data analyst's job,
- 17:42but it won't eliminate it.
- 17:45Anyone who knows data, understands
- 17:47business, and uses artificial
- 17:49intelligence as an assistant, has a
- 17:51secure place in this market. I am very
- 17:53happy that you have joined me this far
- 17:55in the tutorial. If you would like to
- 17:57take the next step and see what the path
- 17:59to becoming an AI engineer is like, be
- 18:01sure to watch this video. Good luck, and
- 18:03I will see you in the next video.
About this transcript
This page contains the full transcript of یادگیری تحلیل داده از صفر با AI | بدون این ۶ قدم، Data Analyst نمیشی! by Hozhi, generated from the public captions YouTube serves with the video. The transcript has 2,853 words across 471 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.