Congrats, you got your first data analyst job! You may be tempted to work yourself ragged, taking every data analytics course, tackling unnecessary data projects, and saying “yes” to everyone. Take a step back and breathe.

To help you get off to a smooth start, I’ll share the top mistakes I made when I began my first data analyst job. Learn from my experience and avoid these mistakes!

Mistake #1: Taking too many data analytics courses at once.

When starting as a data analyst, I wanted to learn as much as possible, so I signed up for classes in everything — Power BI, Excel, Python, R, you name it! But I went overboard.

By taking in too much information at once, I didn’t absorb anything. It was all in one ear, out the other.

What to do instead: Take breaks to apply what you’ve learned in courses to actual, real-life datasets, or nothing will sink in.

Instead of trying to learn every data analytics tool, focus on what you use most day-to-day. For example, if you mainly work in Excel, take Maven Analytics courses in Excel and apply those concepts to real-life datasets.

Mistake #2: Not employing the 80/20 principle.

The 80/20 rule, or the “Pareto Principle,” says that 20% of your efforts often lead to 80% of the results. I wasn’t aware of this principle when I started my job. Instead, I tried to satisfy everyone and treat all tasks as equally important.

This is a recipe for burnout!

What to do instead: When given a task, ask yourself where this fits into your list of priorities. Is it high, medium, or low priority? Which 20% of reports are the most important? Focus your time and energy there first.

Also, be aware of which 20% of stakeholders are the most important. Who do you need to respond to ASAP, and who can wait?

Since you’re starting, you may not know which projects are most important. That’s okay. Regularly check in with your boss about their priorities: “This is my list of tasks. I think Project X is most important right now, but can you confirm?”

Mistake #3: Jumping into analysis without understanding the main business question.

I can’t tell you how often I’ve been asked to perform a “quick analysis” because someone wants to “see the data.” This request is way too vague! I struggled with these vague requests because I needed additional information — what metrics were they interested in? What timeframe did they have in mind?

But most importantly, I didn’t understand the business question.

Before starting any analysis, you need to understand the main business question, or you will waste time and energy going down rabbit holes, answering questions your stakeholders don’t care about.

What to do instead: When you receive a data request, always ask, “What is the main business question?”

Even better, ask them to fill out a data request form that asks for this information.

These are more questions to clarify a vague data request:

• What specific information will help the most?
• What data fields do they need?
• What time frames are they interested in?
• What does “success” look like for this analysis? Are there specific thresholds they’re hoping to meet?
• What business decisions are they trying to make based on this information?

Mistake #4: Setting unrealistic expectations for my first week/month/year, and being disappointed in myself when I didn’t reach them.

Fair warning — when you start your job, people may expect you to create an in-depth dashboard right away, and come up with groundbreaking insights.

Unfortunately, this probably won’t happen. Before you create a dashboard, you’ll need to learn everyone’s name, get set up on a laptop, connect to the proper datasets, and understand what the variables even mean in the business context. This process can take months (it did for me.)

The first year of my data analyst job was by far the most challenging. I was constantly learning and felt inefficient and useless. I beat myself up for not learning faster and thought I was “failing” — that I didn’t have what it took to do the job.

What to do instead: Don’t beat yourself up like I did! When people have unrealistic expectations, explain that you’re getting up to speed. Take PTO regularly to chill out at home and turn off your brain. Stay engaged with hobbies and loved ones outside of work.

It’s hard to believe, but your knowledge will increase over time, and you will understand so much more as you continue in your role.

Mistake #5: Taking what others say too personally.

When you’re overwhelmed, it’s easy to take what others say personally. I struggled with negative feedback, or what I perceived as negative feedback.

Sometimes I introduced new ideas that weren’t taken well by people on my team, or I didn’t create a dashboard as quickly as they wanted. These perceived criticisms weighed on me, even though I don’t think anyone meant to hurt my feelings.

This reinforced my worry that I wasn’t cut out for the job.

What to do instead: When playing a mental movie in your head of what happened, stick to the facts and what they actually said, instead of what you perceived. If you force yourself to stick to the facts and take deep breaths, you’ll find that most of the time, people mean well and do not mean to hurt your feelings.

Practice deep breathing to settle your heart rate before responding. I like to practice 4–7–8 breathing (inhale for 4 seconds, hold for 7, exhale for 8). There are tons of meditation apps and videos you can follow along with, or you can just count the numbers in your head. Find a mindfulness technique that works for you, even if it’s just to take a deep breath before responding.

And remember: how people react to you doesn’t have to do with you. Everyone has their own lives, and they’re probably stressed and overwhelmed by something else. Everyone is the center of their own lives.

Final Thoughts

It’s tempting to impress everyone on your team by trying to do everything — create a perfect dashboard right away, clean all the data in the system, and come up with earth-shattering analyses.

Be kind to yourself during this difficult transition to a new role. I know it’s a bit cliche at this point, but make a point to celebrate small wins.

Did you go to a company lunch to meet everyone, even though you hate socializing? Good for you!

Did you figure out how to clean an Excel spreadsheet? Good for you!

Technical work is challenging and there is a steep learning curve. Take things slow and remember to take lots of breaks!

Further Reading

“Work smarter, not harder: Using the 80/20 principle in data analysis,” by Margaret Efron | Learning Data Maven Analytics

“1 Year In: My lessons from being a data analyst”, by Margaret Efron | Learning Data Maven Analytics

“5 Things I Wish I Knew Before My First Job as a Data Analyst,” by Margaret Efron | Learning Data Maven Analytics

The contents of external submissions are not necessarily reflective of the opinions or work of Maven Analytics or any of its team members.

We believe in fostering lifelong learning and our intent is to provide a platform for the data community to share their work and seek feedback from the Maven Analytics data fam.

Submit your own writing here if you’d like to become a contributor.

Happy learning!

-Team Maven

5 Mistakes I made at my first data analyst job was originally published in Learning Data on Medium, where people are continuing the conversation by highlighting and responding to this story.

ChatGPT is a great tool when you’re faced with an awkward workplace situation and aren’t sure how to respond. I often ask ChatGPT to help me draft emails.

If you’re procrastinating on an email because you don’t know what to say, it’s helpful to work from something, not a blank page!

That said, don’t copy ChatGPT’s output verbatim. It tends to be vague, wordy, and lacks your personal touch. Plus, it doesn’t know the people you’re talking to or the exact tone you need.

Read on for ways I use ChatGPT in awkward workplace situations.

Negotiate for more work-from-home time.

If you want to ask your boss for more workplace flexibility, ChatGPT can help you draft a professional email.

My prompt: “I want to ask my boss for more work-from-home time. Can you draft an email to help me with this request?”

ChatGPT gave me a useful template, including placeholders for details like how long I’ve worked at the company and the tasks I perform better remotely — e.g. those requiring greater concentration.

To edit this response, I’d ask ChatGPT to condense it into four sentences and make the request more direct — for example, increasing remote work from two to four days per week.

Clarify what a stakeholder really wants from their request.

I often get vague data requests I don’t know how to answer, like “Can you pull the data for me?”

Instead of going down pointless rabbit holes, I asked ChatGPT for follow-up questions to get to the heart of the request. This saved me and the stakeholders time.

My prompt: “What are some ways to ask about the ‘business purpose’ of a data request when a coworker asks me to find data? I want to know how this info will be used, what numbers will be considered ‘good’, etc.”

ChatGPT gave a list of useful follow-up questions. I’ve found these especially useful to have in my back pocket:

• “Can you help me understand how this data will be used?”
• “What would success look like for this analysis? Are there any specific metrics or thresholds you’re hoping to see?”
• “What is the goal or decision you’re looking to support with this data?”

These questions make sure I deliver meaningful insights instead of a confusing series of numbers!

Politely decline a meeting invite.

Declining a meeting invite is awkward, yes — but it’s necessary to manage your time effectively.

I often ask ChatGPT to draft emails for these situations, whether I need to decline a meeting or request a last-minute time change.

My prompt: “Can you help me draft an email declining a meeting invite?”

ChatGPT’s initial response was too long and didn’t sound natural, but it provided a starting point. To make it more direct, I used iterative prompts like:

“Can you shorten this to 3 sentences?”

This revised email is much better! It’s concise, yet still professional and firm. If I want to add more personal touches, I can do so easily.

Write a LinkedIn connection request.

I often want to connect with people on LinkedIn who might not immediately remember me — former colleagues, old college friends who knew me by a different name, etc.

Recently, I asked ChatGPT to draft a connection request for someone joining my William & Mary MSBA cohort. I wanted to sound eager to connect, but not too stalker-ish!

My prompt: “Help me write a quick request for a LinkedIn connection. I’m reaching out to Derek, who is part of the cohort for my upcoming William & Mary MSBA program. Keep the request to 3 sentences.”

ChatGPT’s response was great — it was short, friendly, and referenced our shared connection through the program. It sounds both professional and casual.

Request a letter of recommendation from your supervisor.

It’s awkward asking for a letter of recommendation — but sometimes it just has to be done! ChatGPT can help by drafting an email you can refine with personal details, including specific stories or examples.

In the example below, I asked ChatGPT to draft an email to my supervisor requesting a letter of recommendation for a business analytics program.

My prompt: “Draft an email I can send to my supervisor asking for a letter of recommendation for William & Mary’s Master’s in Business Analytics program.”

Yes, this draft email is awkward, stilted, and much too long — but it gave me a good starting point. I asked ChatGPT to condense the email:

My follow-up prompt: “Condense this to four sentences. Remove the part about the application portal.”

This revised version is more direct and polished, and closer to my desired end goal.

Turn down a job offer.

Oof! Turning down a job offer is awkward, especially when you’ve spent weeks going back and forth with a company eager to hire you.

However, it’s important to be courteous and decline a job offer as soon as possible, and politely as possible, to not waste anyone’s time or burn bridges.

Let’s see if ChatGPT can help.

My prompt: “Draft an email for me to send to the hiring manager at Coinbase. I need to decline a data analyst job offer.”

I’m pleased with ChatGPT’s output. It covered my bases — clearly and firmly declining the job offer while thanking the team for their time and consideration.

To edit this draft, I would ask ChatGPT to shorten it to around 4 sentences, to make it more direct and human-like.

Express condolences.

It’s hard to know what to say when a coworker is grieving. You may want to express sympathy but not know how to start. You may want to be there for your coworker, but not unintentionally offend them.

ChatGPT can provide a thoughtful draft that you can personalize with specific details or adjust for tone. You shouldn’t use its response verbatim, but it can be a helpful starting point.

My prompt: “Help me draft an email expressing condolences for a coworker whose father passed away from cancer.”

I wanted to learn more about how to stay polite and professional when expressing condolences to a coworker, so I asked ChatGPT a follow-up question:

My follow-up prompt: “What are some things to keep in mind when expressing condolences to a coworker? I want to make sure to stay polite and professional.”

I agree with ChatGPT’s tips on expressing condolences, particularly these points:

• Avoid lengthy messages that may be overwhelming during a difficult time.
• Avoid overly religious language.
• Offer support, but don’t pressure them to share more than they’re comfortable with.

These are all good things to remember when you want to support a coworker but not make them uncomfortable.

Final Thoughts

ChatGPT is not a substitute for the nuances of human connection, but it’s a helpful starting point when you don’t know what to say.

If you’re unsatisfied with ChatGPT’s output, use these follow-up prompts to make the writing more human-like:

• “Condense this to X sentences.”
• “Make the tone more informal.”
• “Make this email more direct and to the point.”
• “Make the tone more firm, while maintaining professionalism.”

I hope ChatGPT makes your work life a little less awkward!

How I Use ChatGPT to Handle Tricky Workplace Scenarios was originally published in Code Like A Girl on Medium, where people are continuing the conversation by highlighting and responding to this story.

As 2024 winds down, I’m reflecting on the Excel productivity tips with the biggest impact on my workflow this year.

My favorite tips from Maven Analytics included the Quick Access Toolbar, keyboard shortcuts, Autofill, Flashfill, and more.

Read on for my top tips to speed up your Excel analysis!

1. Customize Quick Access Toolbar.

If you use Excel frequently, you’ll have your favorite tools you’ll want to access no matter what tab you’re on. Good news! You can set up the Quick Access Toolbar for this.

The Quick Access tools are located above the ribbon. By default, they include Save, Undo, and Redo.

You can customize this toolbar as you wish. If you click on a tool and add it to the Quick Access Toolbar (e.g. Freeze Panes), it’s accessible at the top of Excel, no matter what tab you’re on.

For example, I frequently use Freeze Panes, so I right-clicked and selected Add to Quick Access Toolbar.

Then, I saw the Freeze Panes icon at the top of Excel:

Add whichever tools you want!

To remove tools, right-click and select Remove from Quick Access Toolbar.

2. Use Autofill & Flash Fill.

Autofill and Flash Fill automate repetitive tasks like filling in dates, numbers, and formulas. They standardize your dataset for easier analysis and reduce the risk of manual errors.

And, best of all, these tools save you time from having to type a series like “1, 2, 3, 4…” or “Mon, Tues, Wed…”

To use Autofill:

• Select the cell(s) with the data. In this case, I want to extend the list of Order Numbers from 1001, 1002, 1003 to 1004, 1005, 1006, and so on.

• Hover your cursor over the small square in the bottom-right corner of the last cell (the fill handle.)
• Drag or double-click to automatically fill other cells based on the pattern or data in adjacent cells.

Excel will repeat values, extend a pattern, or apply the same formula across the cell range.

Excel Pro Tip: If you click on the Autofill icon, there are more options for filling in values, including Copy Cells, Fill Series, Fill Formatting Only, Fill Without Formatting, and Flash Fill.

A quick summary of the differences between these options:

• Copy Cells: This copies the selected values exactly as they are into the new cells, including any formatting. This is useful if you want to repeat the same value or set of values (e.g. “Product C,” “Product C,” “Product C”)
• Fill Series: Extends a numerical or date-based pattern across the selected cells. (e.g. If you have the numbers 1, 2, Excel will continue the sequence: 3, 4, 5…)
• Fill Formatting Only: Copies only the formatting of the selected cells to new cells, without changing their existing data.
• Fill Without Formatting: Fills cells with the same data or pattern but does not apply any formatting (ignores cell color, font styles, etc.)
• Flash Fill: Fills cells based on the pattern it detects. Good for text manipulation (e.g. repetitive tasks like joining names or splitting text.)

3. Add a “Favorites” Tab to your Excel ribbon.

Add your most frequently used Excel tools to a “Favorites” tab for easy access.

For example, for every spreadsheet, I freeze the top pane and create a PivotTable. To save time from navigating between different Excel tabs, I created a “Favorites” tab with these tools saved.

Create a “Favorites” tab by following these steps:

• Go to File > Options.
• Click Customize Ribbon.

• Click New Tab.
• Rename your new tab to something descriptive — like “Favorites.”
• Add your favorite tools to your new tab.

• Go to the new Favorites tab:

I added these tools to my “Favorites” tab: Insert PivotTable from Table/Range, Insert Rows/Columns, Delete Rows/Columns, Format, Percent Style, Increase/Decrease Decimal, and Freeze Panes. But you can add whatever you want — that’s the beauty of it!

Excel Pro Tip: Use ALT keyboard shortcuts to open your Favorites tab. For example, press ALT to open ALT keyboard shortcuts, navigate to your Favorites tab, and select your desired step (e.g. Insert Pivot Table.)

Below, I clicked ALT to open the ALT keyboard shortcuts.

Then, I pressed Y1 to open the Favorites tab.

It may not seem like it right now, but trust me — this is a game-changer!

4. Use Ctrl Keyboard Shortcuts.

Ditching the mouse will save you so much time! Below is just a list to get started.

Here are my favorite Excel Ctrl shortcuts:

• Ctrl + arrow keys: Navigate to the edge of the data region in any direction.
• Ctrl + Home: Go to cell A1 (helpful for large datasets.)
• Ctrl + End: Go to the last cell containing data (again, helpful for large datasets.)
• Shift + arrow keys: Select cells in a direction.
• Ctrl + Shift + arrow keys: Extend the selection to the edge of data in a direction.
• Ctrl + T: Create a table using the data you selected.
• Ctrl + Shift + L: Activates filters.

Don’t memorize all these shortcuts at once. Instead, focus on shortcuts for actions you perform most often and add to the list over time. I learned these shortcuts by having a post-it note by my desk.

5. Use ALT keyboard shortcuts.

Clicking ALT opens a whole world of Excel keyboard shortcuts, so you can use your favorite Excel tools without reaching for your mouse.

For example, I want to insert a PivotTable for the dataset below, using only ALT keyboard shortcuts.

To use ALT keyboard shortcuts:

• Press ALT. Review the keyboard shortcuts in the top ribbon.

• I want to insert a PivotTable, so I press N to go to the Insert tab.

• I press V to select PivotTable from the Insert tab.

• I select T to insert a PivotTable from a Table/Range.

These ALT keyboard shortcuts can be used in Excel to do anything you would do with a mouse.

6. Create PivotTables for quick analysis.

When analyzing Excel spreadsheets, I create PivotTables to review data trends. PivotTables allow you to quickly and easily filter, group, and summarize data without memorizing formulas.

PivotTables, in my opinion, should be required learning for any office professional using Excel — not just data analysts!

To set up a PivotTable:

• Make sure your data is in tabular format, with clearly labeled column headers. Each column should contain one type of data only (e.g. dates, numbers, or text.)
• Select the data range and go to the Insert tab > PivotTable > From Table/Range.

• Choose where to insert the PivotTable (either in a new worksheet or an existing worksheet.)

• Use the PivotTable Fields Pane to drag fields into the Rows, Columns, Values, and Filters.

For example, I created a PivotTable of the authors’ genders by placing Author Gender in Rows and Number of Books in Values.

• Select your desired summary function (e.g. sum, average, count, max, min) by clicking on the fields in the Values area and selecting Value Field Settings.

• To re-label the column headers, double-click on them, and type in the new column name (e.g. I changed “Row Labels” to “Author Gender.”)

This is just one simple example of a PivotTable. I highly recommend playing around with different fields in your PivotTable analysis and using the % of Column Total under Value Field Settings to see the % each category makes up of the whole 100%.

Below, I added % of Column Total to see the percentage of books written by men vs. women.

7. Add sparklines for quick insights.

Sparklines are tiny charts within a single cell, perfect for visualizing data trends at a glance. This is especially helpful for seeing patterns over time (e.g. stock prices, sales trends, tracking spending.)

To add sparklines:

• Highlight the data you want to represent with a sparkline. For example, I want to compare 2023 and 2024, so I highlighted those cells.

• Go to the Insert tab > Sparklines and select the type of sparkline you want to use. In this case, I chose a line chart to show change over time.

• Choose the data you want and where you want the sparklines to go.

• Analyze the sparklines.

The sparklines make it easier to see, for example, that all genre numbers increased between 2023 and 2024 except for science.

Excel Pro Tip: To make the sparklines even more powerful, couple them with conditional formatting to highlight increases or decreases.

8. Use conditional formatting to highlight data outliers.

Conditional formatting highlights trends or outliers in your data. For example, use conditional formatting to highlight cells above or below a certain value, highlight the top or bottom 10% of values, and more.

In the example below, I used Conditional Formatting to highlight all orders over $100.

To use conditional formatting:

• Select the cells to analyze.
• Go to the Home tab > Conditional Formatting.

• Set up Conditional Formatting for what you want to see, e.g. all orders over $100 (Greater Than):

• Review the output. The orders over $100 are now highlighted in red.

To adjust the conditional formatting rules, go to Conditional Formatting > Manage Rules for a list of all rules applied to your dataset.

In the Manage Rules dialog box, add, edit, or delete rules as needed.

9. Use color scales to see trends.

Quickly visualize data trends with color scales. Color scales apply a spectrum of colors to cells based on their values (e.g. green for high values, yellow for mid-values, red for low values.)

To apply color scales:

• Highlight the cells you want to format.
• Go to the Home tab > Conditional Formatting > Color Scales.

• Choose a color scale from the gallery (e.g. green-yellow-red.) Note: See how the color scale will be applied by hovering over it before making a selection.
• In the example below, I chose a green-yellow-red color scale to highlight the most popular genres in green, the mid-tier genres in yellow, and the least popular genres in red.

Now, I see the most popular genres were self-help, romance, and graphic novels (in dark green) and the least popular genre was science.

10. Use Go To Special to replace blank cells.

To quickly replace blank cells with a certain value (e.g. 0), use Go to Special.

To use Go to Special:

• Highlight the range where you want to replace blanks. For example, the chart below has blank cells I want to replace with 0.

• Open Go to Special by pressing Ctrl + G.
• In the box that appears, click Special.

• Select Blanks.

• Now, all blank cells in the selected range will be highlighted.
• Next, type 0 in one of the highlighted cells, but don’t press Enter yet.
• Press Ctrl + Enter to fill all the selected blank cells with 0.

Excel Pro Tip: Use Go to Special for other purposes, e.g. formatting blank cells with a fill color or highlighting cells with formulas.

Final Thoughts

Start 2025 strong by mastering these Excel productivity tips!

If you want to perfect your Excel knowledge in the new year, take the Maven Analytics courses linked below and follow along with your own datasets.

This is a fun way to analyze your personal data — for example, your Amazon orders, your music downloads, your FitBit data, and more.

And remember, you can always ask ChatGPT for help along the way!

Further Learning

“Top Excel Shortcuts You Should Know” — Maven Analytics by Margaret Efron

Maven Analytics course: Excel Pro Tips — Productivity

Maven Analytics course: Excel Formulas & Functions

The contents of external submissions are not necessarily reflective of the opinions or work of Maven Analytics or any of its team members.

We believe in fostering lifelong learning and our intent is to provide a platform for the data community to share their work and seek feedback from the Maven Analytics data fam.

Submit your own writing here if you’d like to become a contributor.

Happy learning!

-Team Maven

Top Excel Productivity Tips I Learned in 2024 was originally published in Learning Data on Medium, where people are continuing the conversation by highlighting and responding to this story.

ChatGPT is a great resource for learning any new programming language, including R. R is an open-source, free programming language specifically built for statistical computing, including regression, time series, clustering, and hypothesis testing.

Are you new to R? That's no problem! ChatGPT can guide you along the way.

Below are my favourite ways to use ChatGPT to learn R, but these prompts can be adjusted to learn any new programming language, including Python.

Learn basic key functions in R.

When starting, you may want to practice basic functions so you don’t get overwhelmed. Ask ChatGPT:

“What are some key functions I should learn for R? I’m a beginner.”

ChatGPT created a long list of basic R functions, sorted into categories: Basics, Working with Data, Data Structures, Loops and Conditional Statements, and more. I could easily browse the list and see which functions I wanted to tackle first.

For example, the Basics section included print() to display the output in the console, c() to combine values into a vector, seq() to create a sequence of numbers, and rep() to repeat values.

I practiced these basic functions in R Studio, and they all worked correctly:

You can also ask ChatGPT to give specific functions you want to practice:

“What are the basic R functions I should learn for understanding a new dataset that’s been given to me?”

ChatGPT gave me a list of helpful R functions for understanding the dataset structure (including str() , dim() , head() , and tail() ), summarizing data (including summary() ), checking data types, and more.

It also provided a workflow example with comments explaining each of the functions. I saved this example to apply it to new datasets in the future.

Learn how to perform a certain action in R.

Sometimes, you won’t remember the needed function to get your desired output. In this case, ask ChatGPT:

“In R, how can I check the variable type for the age variable in data frame ‘gss.2016’?”

ChatGPT listed different functions I could use, but I found the `class()` function was what I needed.

I used this function: class(gss.2016$age)and got the answer [1] “character”, as you can see below, meaning that the age variable is stored as text.

Troubleshoot R code.

When running R code, you will run into errors — it’s just a fact of life! Ask ChatGPT for help and include the following info:

• The purpose of your code (“I’m trying to…”)
• The code you used (“This is the code I used…”)
• The error you received (“This is the error I received…”)

My prompt: “Can you help me figure out why I’m getting an error in R? [List purpose, code, and error message]”

ChatGPT found my mistake — there was a syntax issue with my code. The last mutate() function did not include the necessary pipe %>% before it:

I ran this corrected code through R Studio, and it worked as expected. This saved me hours of headache trying to debug the code!

Ask what a function does.

If you don’t remember what a function does, ask ChatGPT:

“What does the fread function do in R?”

ChatGPT broke down the key features of fread, the syntax, benefits, and when it should be used:

This is helpful if you’re trying to decide which function to use or trying to remember what your code does.

Figure out what your R code does.

If you’re reviewing code you wrote a while ago or code from your course materials, you may not understand what it does when you first read it.

In this case, ask ChatGPT for a step-by-step breakdown.

My prompt: “What does this R code do? “gss.2016$age [gss.2016$age == “89 OR OLDER”] <- “89”””

Check your understanding.

If you think you understand how your code works but aren’t sure, write what you *think* is happening, and ask ChatGPT to correct you.

For example, I wanted to make sure I understood the function na_if() , so I asked ChatGPT:

My prompt: My understanding is that the function na_if() checks for a certain variable, e.g. to see if the item is “DK”, and changes it to “NA” if so. For example, if I use the na_if function — na_if(x = grass, y = “DK”) — , then all items listed as “DK” will be coded as “NA” and treated as missing. Is that correct?

ChatGPT confirmed that my understanding of the na_if() function is correct, and gave key points to keep in mind. Now, I understand the function even better than before!

Ask for an example of how your R code is applied.

Sometimes, even when you understand what code is supposed to do, you want a sample dataset to see the manipulation happening to the numbers.

After presenting ChatGPT with the code above changing “89 or older” to “89”, I asked how this code would change the data.

My prompt: “Can you give an example of how this would work in the data? : “gss.2016$age [gss.2016$age == “89 OR OLDER”] <- “89””

I loved comparing the old and updated datasets in this simple example. The “key takeaways” at the end were also helpful.

Learn R studio shortcuts.

One of my favorite ways to use ChatGPT is to learn keyboard shortcuts!

My prompt: “What are some basic PC shortcuts for R Studio? For example, to run is Control-Enter.”

ChatGPT gave an excellent list of commonly used PC shortcuts for R Studio:

Make an R programming cheat sheet.

I wanted to make an R keyboard shortcut cheatsheet to have at my desk for reference, so, following the above prompt, I asked ChatGPT:

Follow-up prompt: “Make a printable cheatsheet.”

Note: If you want to make a cheatsheet of something else (e.g. basic functions for analyzing a new dataset), you can do that as well — just specify what you want in your prompt.

When I clicked on the cheatsheet link, I downloaded a PDF with the keyboard shortcuts that ChatGPT listed. I printed the sheet for future reference.

Final Thoughts

ChatGPT is an excellent sidekick for learning new data analytics or programming software, including R Studio, Python, SQL, and more. I always have a tab open with ChatGPT when I’m struggling through new code.

If you’re interested in using ChatGPT to help you learn new systems, I recommend my articles:

“5 Ways I Use ChatGPT to Level Up my SQL Skills” by Margaret Efron with Maven Analytics

“13 Ways I Use ChatGPT as a Study Buddy in my Data Science Course” by Margaret Efron with Code Like A Girl

Happy learning!

Use ChatGPT to Become an R Programming Wiz! was originally published in Code Like A Girl on Medium, where people are continuing the conversation by highlighting and responding to this story.

Imagine this common scenario: You’re a data analyst at a company, and your boss asks you to perform a “quick analysis.” But this request is too vague! You don’t know which metrics she’s interested in, the timeframe she has in mind, or even the specific question she’s trying to answer.

Instead of going down random rabbit holes to figure out what they want, ask your stakeholders clarifying questions.

Asking the right questions upfront prevents wasted time and ensures you give your stakeholders what they need.

Below are my favorite clarifying questions when given a vague data request.

1. What is the main business question?

Before starting any analysis, you need to know the main business question, as it guides the data you will collect.

To clarify the business question, ask your stakeholders!

Here are some business questions they may give you in response:

• Did we meet our KPIs this year, and if not, why?
• How does sales performance this April compare to last April?
• What’s been the impact of the New Year’s Eve marketing campaign?
• Who are our most valuable customers?

If the response is still unclear, you may need to ask more follow-up questions.

2. What specific information will help you the most?

Ask them to be as specific as possible — what data fields do they need? What time frames are they interested in? If they’re interested in a particular time frame, do they want to look at year-over-year or month-over-month trends? If so, how far back should you go in the analysis?

For example, if you work for a pet food company and they want to analyze the success of a “Bark the Halls” email campaign, they may ask for these metrics:

• Open Rate: Percentage of recipients who opened the email.
• Click-Through Rate (CTR): Number of recipients who clicked on links in email.
• Unsubscribe Rate: Number of recipients who opted out after receiving the campaign email.

3. What does “success” look like? Are there specific targets or thresholds you hope to meet?

Knowing what constitutes “success” is crucial for interpreting data, and determining if you’re delivering “good” or “bad” news.

For example, I once analyzed an email campaign and found a 35% open rate, which I initially thought was poor. However, after talking with our marketing manager, I learned that anything above 10% was a huge success!

This understanding also led to helpful follow-up questions, including:

• “Why was this campaign so much more successful than previous campaigns?”
• “What made it different?”
• “Can we replicate this in the future?”

4. Have you seen a report like this before? If so, can you send it to me for reference?

I always ask this question — it’s a huge time-saver! If your boss can find a similar report from a month ago, it will only take them a few minutes to share but saves you days of having to recreate the wheel.

Even if the report isn’t exactly what they want, it gives you an idea of what they’re looking for, so you don’t have to start from scratch.

5. Who is the audience for this data? Which teams or stakeholders will be reviewing it?

Understanding your audience is key to determining the right format for presenting results. Executives may want a PowerPoint with charts, while the IT department might require the underlying code and calculations.

Don’t waste your time creating a pretty dashboard if that’s not what they need!

Knowing your audience also helps you anticipate potential questions. For example, if this analysis is for the marketing team, expect questions about campaign performance, audience engagement, and conversion metrics.

Final Thoughts

Ask questions throughout the data analysis process to keep you and your stakeholders on the same page.

Unfortunately, data analysts aren’t mind readers, so it’s important for your boss to clearly communicate their needs before you begin data wrangling.

Bonus tip: To save even more time, share a sample of your analysis with stakeholders along the way.

For example, instead of spending 8 hours on Project X only to realize you’ve been focusing on the wrong variables, share a partial analysis early on. This way, if you’re not on the right track, you can course-correct before investing more time.

The contents of external submissions are not necessarily reflective of the opinions or work of Maven Analytics or any of its team members.

We believe in fostering lifelong learning and our intent is to provide a platform for the data community to share their work and seek feedback from the Maven Analytics data fam.

Submit your own writing here if you’d like to become a contributor.

Happy learning!

-Team Maven

5 Questions you need to ask before taking on any data project was originally published in Learning Data on Medium, where people are continuing the conversation by highlighting and responding to this story.

I’ll make a bold claim:

proper boundaries are the most essential skill for maintaining your sanity in the workplace.

Remember: People take as much as you’re willing to give. This is human nature. Often, they won’t know they’re pushing you to the breaking point unless you tell them.

As Melissa Urban, author of The Book of Boundaries: Set the Limits that Will Set You Free says:

A boundary isn’t about controlling someone else or telling them what to do. Boundaries are a limit around what you will or will not do in order to keep yourself safe and healthy.

Telling people what you are or aren’t willing to do can lead to uncomfortable conversations, so it’s helpful to have scripts for these situations.

Below are times I set boundaries at work with great success. Each time I set a boundary, I felt stronger and better able to stand up for myself.

1. Decline a Meeting Invite

Do an inventory of your upcoming meetings. Is your presence required, or are they wasting your time? Do they lead to productive insights?

If you don’t think a meeting will be productive, decline the invite or at least shorten it.

My Experience

I decline meeting invites if I don’t have anything to add. For example, we had unnecessary meetings with our vendors. If I didn’t have anything to contribute, I emailed:

“Thanks, but I don’t have any questions right now, so will decline this meeting invite. If anything comes up, I will email you.”

I haven’t received pushback for declining meetings. Most of the time, people are relieved to have extra time in their day or don’t think my attendance was crucial.

Example Scripts:

“Thanks for thinking of me. Unfortunately, I have a conflict and won’t be able to attend. Please let me know if there’s anything crucial I missed.”

“Thanks for the invite. I won’t be able to make it, but I will catch up with Jessie about what I missed.”

“I’m busy focusing on XYZ projects, so I won’t be able to make it. Thanks for thinking of me.”

“I’m busy with ABC project, so will only be able to attend the first fifteen minutes.”

2. Decline a Project

How often does this happen: a coworker asks you to run a data analysis, but your to-do list is a mile long already?

My experience

My coworkers often don’t consider the other tasks on my plate. This is to be expected — they’re focused on completing their job, just as I’m focused on mine.

They don’t know what I’m up to unless I explicitly tell them. To counteract this, I remind my coworkers of other data requests in my queue:

“I’m currently working on a project for Angela, so my plate is full.”

Reminding my coworkers of my other tasks has many benefits. It makes them think twice before asking me for an unnecessary data request during busy times, helps them value my time, and gives them a better sense of my competing priorities.

Example Scripts:

“Thanks for thinking of me, but my plate is full. Maybe you can ask Christie?”

“I appreciate your thinking of me, but this is outside my scope.”

“I’m sorry, I’m focusing on XYZ and don’t have time to take on this data request right now.”

“I have four other data requests that I’m working on for you. Is this more important? If so, what should I put on the back burner?”

“Thank you for thinking of me. This project doesn’t feel like a good fit, so I’ll pass this time.”

3. Confront an Unrealistic Workload

Even though you decline some work projects, there may still be too much to handle. Your job may stretch you too thin, with too many requests from too many people.

In this case, check in with your boss about their priorities. Don’t waste too much of your time and effort on something that isn’t valuable to you or your supervisors.

My Experience

I was overwhelmed, so I listed my tasks and ranked them in order of priority. Then, I asked my boss if I had correctly prioritized the tasks, and asked about her current and future priorities.

Example Scripts:

“I have a lot of tasks on my plate right now, so I want to check in with you to see if I’m prioritizing what I should. Just to be clear, XYZ is our top priority right now, correct?”

“I see that you sent in another data request today. I’m still working on your previous data request. Is this new request more important?”

“I have these tasks I’m working on today: A, B, C. If I could only finish one today, what would you want me to focus my time on?”

“Looking ahead, I know that winter will be busy with project A. Keeping that in mind, I will table project B for now.”

4. Decline Alcohol

There are many reasons you may not want to drink with coworkers. The good thing is you don’t have to explain your reasons. You can just politely decline!

You’d be surprised at how often a casual “No, thank you, I’m not drinking” goes over fine. The less of a deal you make of it, the less of a deal others will make.

My Experience

I’m not a drinker, so whenever alcohol is offered at a work party, I politely decline, saying,

“Thanks, but I’m not drinking today.”

I’ve experienced very little pushback, except for a Christmas office party, to which I put down the drink they handed me and didn’t touch it.

Example scripts:

“No, not today, thanks.”

“I’m not drinking right now, thanks for the offer.”

“I’m good sticking to sparkling water for the rest of the night, thanks.”

5. Deal with Frequent Interruptions

If you’re like me, frequent workplace interruptions drive you batty. Just when you’re in the zone working on an analysis, Bob will stop by asking about your day. It’s not that you don’t like Bob, but the timing is bad! It takes you ten minutes to refocus, and then you’re grumpy.

My Experience

Coworkers constantly stopped by my office to interrupt me, even if my door was closed.

I developed these strategies to make my “deep work” time clear to coworkers:

• Wearing headphones and communicating that “when my headphones are on, I’m in a meeting/focusing/etc.”
• Leaving the office to work in a separate space for privacy.
• Being honest and setting boundaries with my office mate, e.g. “I’m about to jump on a Zoom meeting, so I will go to Room XYZ to focus.”

Example scripts:

“Oh, I’m in the middle of another data project right now. Can I swing by your office after lunch?”

“I have my headphones in because I’m trying to focus. Can you email me and I’ll get back to you later?”

“I’m in a meeting, that’s why my door is closed. Can you email me instead?”

“Hi, I’m busy working on Angela’s data request. Can we talk about this at our next meeting?”

6. Redirect Unwanted Comments About Weight or Diet

Let’s face it. Talking about weight or diet is uncomfortable. However, your coworkers may not know what makes you uncomfortable unless you tell them.

For example, if they comment on your weight loss, they may not know this makes you feel bad. Weight loss is celebrated in diet culture, so they might think you’ll appreciate their comment.

My Experience

Once, a work colleague commented on my weight loss. I took a second to check in and figure out what to say. Then, I said:

“Oh, I don’t feel comfortable talking about my weight at work.”

I redirected the conversation to something innocuous, like the weather.

What happened? It went fine! She apologized, and it never happened again.

Example Scripts:

“I know you mean well, but I’d rather not discuss my weight. That’s a personal subject.” (Then, redirect the conversation to something else.)

“Can we please stop talking about my lunch?” (Redirect the conversation to something else)

“I know you’re trying to help, but I have a health plan I’m working on with my doctor, and I don't want to discuss my diet right now.”

“No, thanks, I’m not hungry right now.”

“That looks delicious, but not tonight.”

Final Thoughts

If you have identified situations that make you uncomfortable, congratulations! Realizing that something isn’t working is the first step.

Next, set boundaries using kind and clear language. Remember: in most cases, your coworkers mean well and the last thing they want is to cause you undue stress.

When setting boundaries with coworkers, keep in mind that they probably mean well and don’t want to stress you out. Approach from a stance of goodwill.

Speaking up for yourself is uncomfortable, but it’s necessary. It helps you be a better worker and tells others how to support you.

Be kind to yourself and your coworkers, and set that boundary!

Further Reading

The Book of Boundaries: Set the Limits That Will Set You Free, by Melissa Urban

Maven Analytics Learning Data: How to Stop Being a People Pleaser at Work: A Data Analyst’s Guide by Margaret Efron

Career Contessa: 10 Signs You’re a Chronic People Pleaser — and How to Stop

Towards Data Science: How To Say No to Useless Data Science Projects and Start Working on What You Want

Forbes: 10 Ways to Stop Being a People Pleaser at Work

6 Powerful Ways to Set Boundaries at Work and Thrive Now was originally published in Code Like A Girl on Medium, where people are continuing the conversation by highlighting and responding to this story.

Have you heard of the 80/20 rule, or the Pareto Principle? It says that roughly 80% of the effects come from 20% of the causes.

In most cases, a small percentage of efforts drive most of the results. Let’s apply this rule to data analysis, and work smarter, not harder!

Why is the 80/20 rule useful? It lets you focus on the few tasks that generate the most value for you and your organization. This saves time, increases efficiency, and makes you more useful at work.

If I could go back in time and tell myself something at the beginning of my data journey, it would be to use this principle.

Below are my favorite ways of applying the 80/20 rule to data analysis, so I work smarter, not harder.

1. Prioritize high-impact tasks.

What 20% of tasks lead to 80% of my results?

Focus on tasks with the most significant impact — like preparing reports for your boss or addressing time-sensitive data issues. Delegate or postpone lower-impact tasks.

Struggling to figure out which tasks are most important? Consider which tasks have the most visibility and impact. Consult your supervisor. Don’t be afraid to ask stakeholders, “What’s your top priority right now? If I could complete just one task today, what would it be?”

2. Master the data platform you use the most often.

Which data platforms do I use 80% of the time?

Since I spent a lot of time cleaning data in Excel, I focused first on learning the functions for cleaning data. This kept me from getting overwhelmed by trying to learn all the data tools at once.

For example, I focused first on practicing TRIM , CLEAN , PROPER, and CONCAT . I watched Maven Analytics videos about these functions and followed along with my datasets.

3. Use learning methods that YOU find most helpful.

Which 20% of learning methods yield 80% of my improvement as a data analyst?

Everyone has different preferred ways of learning, whether from YouTube videos, structured learning platforms, personal data projects, graduate degrees, books, or something else!

Experiment and find what works for you. I found it most helpful to watch Maven Analytics videos on the data tools I use 80% of the time and follow along with my data projects.

4. Focus on the most common data quality issues.

Which 20% of data quality issues lead to 80% of my cleaning?

If you deal with messy data, you’ve likely noticed certain errors coming up again, and again…and again. Tackle those common problems first.

Make a data-cleaning checklist of the most common errors you encounter. Keep a running list by your desk during your next project.

Then, set up Data Validation rules to prevent these errors in future data entries. For example, in Excel, you could set:

• Date ranges: Ensure dates are entered in the correct format.
• Dropdown lists: Use dropdowns for consistent data entry (e.g. country names, product names.)
• Custom rules: Use formulas to set custom rules (e.g. so that a value is in a custom range or meets specific criteria.)

5. When taking a course or reading a book, focus on the chapters most relevant to your data analysis and skim the rest.

What 20% of this book is most relevant to my job?

For example, I read a data storytelling book with a chapter on pie charts, which I don’t use for my work. Instead of reading that chapter, I skipped it and read the chapter about line graphs, since my boss prefers to see year-over-year impact.

6. Reduce unproductive meetings.

Which 20% of my meetings are most productive? For the remaining 80%, cancel them or make them shorter.

For example, my vendor check-in meetings were unproductive, as I had nothing to contribute. I canceled them, freeing time for more constructive work.

Here are some ideas if you can’t cancel the meeting:

• Request the meeting be shortened.
• Ask someone to catch you up afterward.
• Propose an alternating attendance schedule with colleagues.
• Let the organizer know you may need to leave early due to other priorities.

7. When creating a dashboard, highlight the most important stats for stakeholders.

Which 20% of data points contribute to 80% of my organization’s insights?

Focus on the most impactful visualizations (e.g. key metrics and trends) to create clearer and more useful dashboards.

For example, when I present data to my stakeholders, 80% of their questions are about year-over-year trends. Now, when I create new dashboards, I focus on year-over-year trends and highlight this information at the top.

8. Master core functions.

What 20% of functions and shortcuts do I use 80% of the time?

I often use Excel PivotTables, so I allocate half an hour a day to Maven Analytics self-paced PivotTable courses. Learning PivotTable tips and tricks has sped up my analysis considerably!

This way, I don’t waste time learning every data trick under the sun, and can immediately apply what I’ve learned to my work.

9. Use shortcuts for common actions.

What 20% of actions do I perform most often? Are there shortcuts that can make this faster?

I listed my most common actions and memorized those keyboard shortcuts. To learn them, I kept a sticky note on my desk with a few new keyboard shortcuts per day.

These are my favorite keyboard shortcuts:

• Ctrl + T to add a new tab.
• Ctrl + W to close a tab.
• Ctrl + Shift + T to open an accidentally closed tab.
• Ctrl + Shift + W to close all tabs.
• Ctrl + D to bookmark a tab.
• Ctrl + Tab to go to the next tab.

I also frequently use Excel, so I listed my favorite Excel actions, and memorized the keyboard shortcuts:

• Ctrl + arrow keys to navigate to the edge of the data region in any direction.
• Ctrl + Shift + arrow keys to extend the selection to the edge of data in a direction.
• Ctrl + Shift + L to activate filters.
• Ctrl + T to turn data into a table.

10. Automate repetitive tasks.

What 20% of tasks take up 80% of my time? Are any of these repetitive? Can any of these be automated — for example, formatting reports, refreshing data, or performing the same cleaning steps?

I spend most of my time cleaning data. If I regularly clean similar datasets, I can automate this using macros or Power Query.

To do this, I recorded a macro in Excel to clean up common issues, including removing extra spaces, formatting numbers, and removing extraneous columns.

Final Thoughts: Look out for yourself and prevent burnout.

As you continue your data analyst journey, I hope you use the 80/20 principle to save time and headache!

Throughout your workday, reflect on how you’re allocating your resources.

Instead of endlessly editing every email, focus on the 20% of emails with the most impact.

Instead of prioritizing all data requests equally, choose the top 2 or 3 tasks with the most impact, and focus on those first.

This is a great way to prevent burnout and practice self-care!

Further reading

PC keyboard shortcuts you need to know | Learning Data | Margaret Efron

Top Excel Shortcuts You Should Know | Learning Data | Margaret Efron

Maven Analytics YouTube Video: Keyboard Shortcuts for Excel in 5 Minutes

The contents of external submissions are not necessarily reflective of the opinions or work of Maven Analytics or any of its team members.

We believe in fostering lifelong learning and our intent is to provide a platform for the data community to share their work and seek feedback from the Maven Analytics data fam.

Submit your own writing here if you’d like to become a contributor.

Happy learning!

-Team Maven

Work smarter, not harder: Using the 80/20 principle in data analysis. was originally published in Learning Data on Medium, where people are continuing the conversation by highlighting and responding to this story.

When I started as a data analyst, I had no idea I would use Excel PivotTables every day. With each new trick I discover, my work becomes faster and more efficient!

Below, I share my favorite PivotTable features, including showing values as a % of column total, summarizing data by sum/average/count, and counting text fields. Follow along to level up your PivotTable game!

1. Show values as % of column total.

The Show Values As options apply calculations to change how values are shown, such as percentages of totals or subtotals, running totals, and ranks. In this example, I show the Total Cost as a percentage of the Column Total.

Below, my PivotTable lists categories and their associated costs, but I want to see the composition — specifically, what percentage of the total cost each category represents (e.g., what percentage was spent on Health?)

To do this, I use Show Values As:

• Go to PivotTable Analyze tab > Design > Grand Totals and select On for Rows and Columns.

• Right-click on a value and select Show Values As > % of Column Total.

Now the PivotTable shows the percentage each category contributes to the total cost, with the entire column totaling 100%.

This is an easy way to analyze the composition and relative share. I quickly see that health products account for 40% of the total cost.

2. Summarize values by Sum, Average, Count.

Summarize Values By controls how data is aggregated in a PivotTable, whether as a sum, count, average, max, or min. This setting is crucial for ensuring that your data is interpreted correctly.

Excel Pro Tip: If a column contains blanks or non-numeric values, Excel defaults to Count when summarizing. It’s often better to change this to Sum for numerical data.

To check how a value is summarized:

- Right-click a PivotTable value and select Summarize Values By. See what is checked. In the example below, the value is summarized by Sum.

• If needed, switch to a different calculation type, such as Average, Max, or Min (e.g., to show Average Gross Revenue, Max Gross Revenue, or Min Gross Revenue).

Another way to check the summary type:

• Click the drop-down arrow next to a field in the Values area (e.g., Gross Revenue).

• Select Value Field Settings:

• Choose how you want the data to be summarized.

Checking your Value Field Settings ensures you analyze what you need to and don’t misinterpret the data. You don’t want to report on an Average by incorrectly calling it the Count!

3. Group numerical values into custom buckets to see the data distribution.

Just like text, you can group numerical fields into custom buckets. This is especially helpful if you want to analyze data distribution or create bins as source data for a Histogram or Pareto Chart.

Using value grouping, I can analyze the count or frequency of observations that fall into specific groups or “bins” of values.

In the example below, I sorted items by price. I want to group the prices into custom buckets to make it easier to see the distribution of orders. Were most orders high-value items, or lower-value items?

To group numerical values:

• Right-click a value in the Price column.

• Select Group.
• Enter the start value, end value, and group size for your custom buckets.

• After grouping, review the output to observe patterns. Most orders are in the $1-$30 range.

• To ungroup, right-click a grouped value and select Ungroup.

4. Count text fields.

The PivotTable Values pane is typically used for numerical fields but can also count text fields.

This is helpful for seeing the number of rows that fall into specific categories, such as movie titles by genre or products by category.

Since you can’t aggregate text fields, use Count Of to display their frequency.

In the example below, I have a PivotTable with product categories and want to know how many products fall into each category.

To do this:

• Drag an additional instance of Category into the Values field.

• Make sure that the Category is summarized as Count Of. (If not, right-click on the value, select Value Field Settings, and set it to Count Of.)
• Review the output. Now I have a new column with the Count of Category.

This helps me see quickly that most products (55) fall into the health category.

Excel Pro Tip: If a numerical field defaults to a count, it usually means values are formatted as text or the source data has blank rows.

5. Click on a value to see the source data.

Let’s say the PivotTable produces a result of “674 books” and I want to see the underlying data that makes up this number.

All I have to do is double-click that value, and a new tab shows the source data behind the number.

In the example below, I have a PivotTable of Goodreads books. The problem is, I don’t have a genre listed for every book. I want to investigate this further!

I click on the value “674” next to “Genre: (blank)” to open a new tab with the full list of books missing a genre.

This helps review blank values, check underlying source data, and ensure that the PivotTable analyses work as expected. If a value looks off, click on it to see the underlying data.

Final Thoughts

Learning PivotTable tricks takes time, so be patient with yourself! I recommend practicing these new features with personal data, like tracking your Amazon orders or analyzing your Goodreads books. Working with familiar, low-stakes data makes learning fun and stress-free.

Then, when it comes time, you can impress everyone at work with your new skills!

Check out the Maven Analytics courses listed below and follow along with your datasets. Practice is truly the only way to learn.

Further Learning:

Maven Analytics Excel Pro Tips: PivotTables course with Chris Dutton

Maven Analytics Microsoft Excel: Pivot Tables & Pivot Charts course with Chris Dutton

The contents of external submissions are not necessarily reflective of the opinions or work of Maven Analytics or any of its team members.

We believe in fostering lifelong learning and our intent is to provide a platform for the data community to share their work and seek feedback from the Maven Analytics data fam.

Submit your own writing here if you’d like to become a contributor.

Happy learning!

-Team Maven

I use Excel PivotTables every day. Here are 5 tricks I use all the time! was originally published in Learning Data on Medium, where people are continuing the conversation by highlighting and responding to this story.

If you’re taking data science courses, you may need a helper on the side… a study buddy. This is where ChatGPT comes in!

Let’s get the disclaimers out of the way: remember that ChatGPT hallucinates, or makes up information, so you need to double-check the output for accuracy. Don’t enter any confidential information. Review your school’s guidelines to ensure you aren’t breaking any rules.

And don’t use ChatGPT to cheat!

This is how I use ChatGPT as my data science study buddy:

1. Learn the difference between joins.

I struggled to remember the different joins, so I enlisted ChatGPT’s help.

My prompt: “Can you describe the difference between these four concepts, using the example of 2 tables, the acquisitions table and the final payments table? ‘Full Join,’ ‘Inner Join’, ‘Left Join,’ ‘Right Join’”

ChatGPT listed joins with examples.

My only critique is that I wish the “visual summary” was visual and not just descriptive. I wanted pictures to accompany the text.

I asked a follow-up question about the difference between a “left join” and a “left outer join”:

My prompt: “So is a left join the same thing as a ‘left outer join’?”

2. Understand basic R functions.

I used ChatGPT to grasp basic R functions since I haven’t worked in R before.

However, you can use ChatGPT to grasp the basics of any programming language, and I’ve found it particularly helpful for SQL (see my article, 5 Ways I Use ChatGPT to Level Up My SQL Skills.)

My prompt: “Can you explain the summary() function in R and why it is useful for summarizing info about elements in a data frame?”

ChatGPT explained how the summary() function summarizes numeric data (by min, 1st qu., median, mean, etc.)

It also discussed why summary() is useful:

• Gives a quick overview of each column.
• Identifies outliers in the dataset, which you may want to examine with further analysis.
• Helps identify skewness or symmetry in numeric data.
• Detects missing values by counting the NA values.

ChatGPT gave me an example of a data frame called df:

I wanted to understand how the summary() function applies to dates, so I asked ChatGPT:

“In R, can you use the summary() function to analyze a date range?”

3. Explain code snippets.

You’ll see new code in your course and may not know how to interpret it. ChatGPT is excellent at breaking down code and explaining what each part does. You can also use ChatGPT to troubleshoot errors when running your code.

My prompt: “Can you explain the syntax in the R code ‘summary(df$DateColumn)’?”

I now understand that df$DateColumn extracts the date column from the data frame, and summary() summarizes the range of dates in that column.

If your courses involve coding, ask ChatGPT what different functions do to see if you’re applying them properly.

4. Ask for practice problems to test my understanding.

I learned about the different joins in my class but didn’t fully grasp the concepts, so I asked ChatGPT for practice questions. It was a fun, low-pressure way to fill in the gaps in my knowledge.

My prompt: “I’m taking a course on data science and want to better understand the difference between full joins, inner joins, right joins, and left joins.

Can you ask me practice questions one by one, wait for my response, and then tell me whether I’ve gotten it right?

Only proceed to the next question once you’ve explained whether I’ve answered correctly.”

ChatGPT asked me questions about the different joins. If I wanted to keep going, I could ask for more questions on any concept or more examples.

5. Create flashcards on specific concepts I need to review.

Does anyone else like using flashcards to quiz themselves? If so, ask ChatGPT to help generate flashcard ideas, and write them on index cards or an online quiz platform. I asked ChatGPT to generate flashcards about join types.

My prompt: “Can you create flashcards for me on these concepts: full join, inner join, right join, left join?”

You can also ask ChatGPT to turn the flashcards into a quiz with specific examples.

6. Generate a quiz with specific examples.

I liked ChatGPT’s flashcards about joins and wanted a full-length pop quiz with specific examples to test myself.

My prompt: “Create a quiz about INNER JOIN, LEFT JOIN, RIGHT JOIN, and FULL JOIN with specific examples.”

ChatGPT created a very helpful 6-question quiz. All the questions were on-topic and included specific examples.

I wrote my answers on my notepad, then told ChatGPT my answers and asked for feedback.

My follow-up prompt: “These are my answers. Please tell me whether they are correct and provide feedback on any incorrect responses: 1)a, 2)b, 3)b, 4)a, 5)a, 6)b.”

When I got an answer wrong, I asked for clarification:

“Can you explain why I got Question 6 wrong? I thought that it is still correct that all employees will be listed in a left JOIN.”

If you want to keep practicing, ask ChatGPT:

“Taking into account the questions I got wrong, can you create a new quiz to solidify those concepts for me?”

7. Summarize lecture notes.

ChatGPT is excellent at summarizing text, including main points from lectures. Before doing so, make sure it’s okay according to your school’s policies and procedures.

My prompt: “Can you summarize my notes on dealing with missing data? [Insert notes here]”

ChatGPT summarized my lecture notes, but I wanted to dig deeper. I asked:

“What would be the pros and cons of these different options for handling missing data?”

8. Refresh math skills.

If you haven’t taken math in a while, ask ChatGPT about new concepts, like continuous and categorical variables, logistic regression, and linear regression. This gets you up to speed without slowing the rest of the class down.

Pro tip: Want to impress your teacher? Look ahead in the syllabus and practice upcoming concepts with ChatGPT before they’re covered in class. This will give you an advantage before the lecture!

Below, I asked ChatGPT how to interpret linear regression results.

My prompt: “What are some basic tips for interpreting linear regression results?”

ChatGPT broke down a linear regression result and added practical tips and examples, including:

9. Ask for a specific example of a math concept.

If you want a specific example, ask ChatGPT:

“Can you provide a sample linear regression interpretation?”

Below, ChatGPT generated a scenario predicting home prices.

ChatGPT provided the regression equation, the results, and the interpretation, which I’ve sampled below. I appreciated how ChatGPT broke down each variable and what it means in the real world:

10. Understand the difference between math concepts.

When asking ChatGPT for a math refresher, ask about the difference between math concepts. For example, if you’ve already covered linear regression and you see logistic regression is next week on your syllabus, your professor will definitely want you to know the differences between them.

Pro tip: Use ChatGPT to understand the difference between the math concept you covered last week (the familiar concept) and what’s coming next week (the new concept.)

My prompt: “Can you explain the difference between linear and logistic regression?”

ChatGPT listed the differences between linear and logistic regression.

It emphasized that the two types of regression have different applications, and gave a comparison summary table:

11. Look ahead in the syllabus to prep for next week and brainstorm new ideas.

I checked my data science class syllabus to see what will be covered next week. The top exercise is “Determine when time series analysis is useful and informative.”

Hmm! If I don’t know about this concept, I can ask ChatGPT:

“When would time series analysis be useful and informative?”

ChatGPT listed examples of when time series analysis may be informative, including measuring cyclical behaviour, assessing the impact of a marketing campaign, measuring the effect of weather events, and more.

It ended with a helpful bulleted summary:

Now, when the concept of “time series analysis” comes up in class discussions, I will already have examples of when it is useful.

12. Ask for book recommendations.

If you still want more, ask ChatGPT for book recommendations. Tell ChatGPT what you want to learn and indicate your knowledge level (beginner, intermediate, or advanced.)

My prompt: “What are some easy-to-understand books about SQL for beginner data scientists?”

ChatGPT listed 8 SQL book recommendations, summarizing why they think the books would help, what I would learn, and who the target audience is.

13. Ask for learning platform recommendations.

You can also ask ChatGPT for website recommendations.

My prompt: “Can you give me website recommendations where I can learn more about SQL?”

ChatGPT recommended many interactive and beginner-friendly platforms, including W3Schools, SQLZoo, Khan Academy, etc.

Fair Warning: ChatGPT doesn’t work perfectly for everything.

ChatGPT hallucinates, and you need to double-check its output for accuracy. Its ability to create images is still hit-or-miss.

I asked it to create images of joins, with bewildering results!

My prompt: “Create an image of this type of join: all info from the acquisitions table will be included, as well as any row from the final payments table with a loan ID that matches one in the acquisitions table.”

I edited the prompt, telling ChatGPT:

“That’s not quite right. Can you just draw a picture of the LEFT JOIN between two circles: Acquisitions table and Final Payments table?”

As you can see, ChatGPT is still learning how to create learning visuals. But since it is improving, I hope it will generate helpful visual summaries in the future!

Further Reading

Learning Data Maven Analytics — “5 Ways I Use ChatGPT to Level Up My SQL Skills” by Margaret Efron

13 Ways I Use ChatGPT as a Study Buddy in my Data Science Course was originally published in Code Like A Girl on Medium, where people are continuing the conversation by highlighting and responding to this story.

Have you heard of the 80/20 rule, or the Pareto Principle? It says that roughly 80% of the effects come from 20% of the causes.

When I learned about the 80/20 principle, it blew my mind. Incorporating the principle into my work helps me focus on what matters, reduce burnout, and stop wasting time.

Below are my favourite ways to use the 80/20 principle at work.

Feel Confident

What 20% of clothing items/makeup/jewelry/accessories make you feel the most confident?

Look at your wardrobe and your beauty counter. Do you use all these items? Do they contribute equally to your confidence at work?

If you’re honest with yourself, the answer is probably “no.”

What clothes make you feel the most powerful? This could be a tailored blazer, a leather jacket, or a silver necklace. (For me, it’s my blazer and silk scarf.)

If you’re going to a job interview or a presentation, now you’ll know what to reach for.

What makeup makes you feel the most put together? This is especially helpful if you’re in a rush. (And if you don’t do makeup, no judgment, just disregard this!)

No worries if you feel 80% put together with mascara and lipstick and you’re running late. Just swipe on those two products and you’re good to go.

Stop Attending Unnecessary Meetings

Which 20% of your meetings are the most productive? For the remaining 80%, cancel them or make them shorter. Stop wasting your time!

Look at your upcoming meetings and identify the 20% most impactful — maybe your boss wants you there, or you’re presenting key insights to the team.

If an upcoming meeting won’t be impactful or productive, cancel it, shorten it, or ask someone to catch you up afterward.

I was initially intimidated and thought I had to attend all the meetings I was invited to. However, once I got over my fear, I found that declining meetings had many benefits. Nothing bad happened, I got time back in my day…and coworkers respected and valued my time since they knew I had competing projects.

If you feel awkward declining meeting invites, try these scripts:

• “Thank you for inviting me. Unfortunately, I have a time conflict. Please let me know if you need my input, and we can connect by email.”
• “Thanks for the invite. My main priority is the XYZ project right now, so I’m afraid I don’t have time to attend. I will touch base with Bella afterward for updates.”
• “Hi, I’m afraid I can’t attend the whole meeting due to conflicts but am happy to attend the first 15 minutes if that would be useful. Please let me know if that works.”

Identify Key Stressors at Work and Tackle Those First

What 20% of stressors cause 80% of your work anxiety?

Track what’s causing your work stress. You can do this through journaling, keeping a notebook at your desk, or asking your friend or partner what you complain about the most. Look for patterns.

Once you have a list of stressors, consider the 20% that cause most of your work anxiety. Are any of these within your control? You may be surprised.

For example, I was annoyed by constant interruptions at work. Coworkers stopped by my office to interrupt me several times an hour, even if my door was closed.

I tackled this issue in several ways:

• Wearing headphones when engaged in deep work.
• Communicating with my office mate about when I needed to focus.
• Leaving the office to work in a separate space for privacy.
• Being honest and setting boundaries with coworkers — e.g. “I’m in the middle of something right now, that’s why my door is closed. Can you email me instead?”

Instead of tackling every stressor I experienced, I saved time and energy by focusing on what made me the most anxious.

Your primary stressor may be a project deadline or a repeating unproductive meeting. Whatever it is, tackle that first. Can you move back the deadline? Can you cancel the unproductive meetings?

Be warned: this may require asserting boundaries. If you have difficulty setting boundaries, I highly recommend The Book of Boundaries: Set the Limits That Will Set You Free by Melissa Urban, which provides hundreds of scripts for handling tricky work and life situations.

Identify What Relaxation Methods Work the Best for YOU (not necessarily for anyone else.)

What 20% of relaxation methods lead to 80% of my stress reduction?

Everyone has different preferences for how they want to relax. Some people feel most relaxed during yoga. Others may prefer getting a lash lift, reading a magazine, talking with a friend, playing ultimate frisbee…the list goes on and on.

What helps you feel the most relaxed? Be creative!

Try incorporating your relaxation habit into your workday. For example, I feel most relaxed when I’m outdoors, so I take a walk outside every day at lunch.

Focus on the Tools that Will Most Impact Your Career Development

What are the 20% of tools will lead to 80% of your career growth?

There’s no need to learn every coding and data visualization platform—that’s a surefire way to feel overwhelmed, burned out, and inefficient.

Instead, think about what tools will help you most in your career. You can clean data in Python, build databases in MySQL Workbench, or analyze data in Power BI. Focus on the one thing you think will improve most.

Ask yourself:

• Which tools will help me the most in my day-to-day work?
• What platforms am I already using? Can I master the core functions for those?
• Which tools are necessary for the next stage in my career?
• What tools do I need to be able to put on my resume?

Final Thoughts

I’ve used the Pareto principle in many ways at work, not just the ways I’ve listed here. Be thoughtful and honest with yourself.

Ask yourself:

• Which 20% of learning methods yield 80% of my improvement as a data scientist?
• What are the 20% of coworker boundaries that, if set, would alleviate most of my anxiety?
• Does this new request align with my 20% of impactful activities? If not, can I decline it?
• Am I wasting time being a perfectionist on the less impactful 80% of tasks? Can I focus on the most impactful 20% of tasks, and not try as hard on the rest?

I hope the 80/20 principle helps you as it has helped me: to reduce work stress, overcome perfectionism, and create more time for the things that truly matter (and let’s be honest — that’s not work).

Further Reading

The Book of Boundaries: Set the Limits That Will Set You Free, by Melissa Urban

Maven Analytics Learning Data: How to Stop Being a People Pleaser at Work: A Data Analyst’s Guide by Margaret Efron

Career Contessa: 10 Signs You’re a Chronic People Pleaser — and How to Stop

Towards Data Science: How To Say No to Useless Data Science Projects and Start Working on What You Want

Forbes: 10 Ways to Stop Being a People Pleaser at Work

Chill Out at Work: Reduce Stress by Applying the 80/20 Principle was originally published in Code Like A Girl on Medium, where people are continuing the conversation by highlighting and responding to this story.