AI Insights

The Most Underrated Skill In Your Successful Data Science Journey

May 28, 2022


article featured image

How to master the skill of asking for help on LinkedIn, via email, and on the job to learn faster, make friends, and take your data science career to the next level.

Asking for help as a data scientist

Seeking help vs. wanting help

Let us let the cat out of the box. Why do we have such a hard time when asking for help? Intrinsically, we all know that asking for help comes with many, even science-proven, benefits such as connecting with others, increasing productivity, maturing our mindset, and making us happier.

Now, we might all have our own reasons for not asking for help, but there are a few common thinking patterns, such as:

  • “I would think less of myself for needing help.”
  • “The problem didn’t seem worth getting help for.”
  • “I wouldn’t know what sort of help is available.”
  • “The problem is embarrassing.”
  • “I don’t like to talk about my feelings.”

And then there is this other issue, the imposter syndrome, which has become a discussion topic among tech influencers like Lex Fridman and many more. Another reason why asking for help can be difficult, especially in data science

The good news is asking (the right way) is not a weakness at all. In fact, it can become one of your biggest strengths.

asking for help data science

Source: Imgflip.com

Here is what you should NOT do

Let’s start with a story you might relate to.

Danny and Pat were solving a Machine Learning problem when Pat challenged Danny that he’ll get a better accuracy score than Danny. Halfway into solving the problem, Pat sees that his code is giving some strange error. He searches on Stack Overflow only to get confused further and realize that he lacks a lot of knowledge. Meanwhile, Danny just completed the challenge and asked Pat about his results. Pat says, “I’m just on it, mate. Just squeezing the most out of the performance!”. Pat just couldn’t admit that he was stuck and asked Danny for help as it would have crushed his ego and made him feel like a loser.

The final result? Danny learned a lot from the challenge while Pat just left it midway being too arrogant to ask for help. Danny somehow gets to see Pat’s code and asks Pat why he didn’t complete the problem. Pat admitted that he couldn’t get through that error and didn’t want to ask him for help.

I couldn’t even reach till where you got! I was stuck and reached out to a friend multiple times who helped me out and only then I was able to finish the problem. You should have asked me once at least!” said Danny. Now, Pat realizes his real mistake and asks Danny how to debug that error and complete his problem.

Result? Pat ended up getting a better accuracy score than Danny.

Three rules to ask for help effectively

Asking for help is a highly underrated skill that most people face when learning, doing something new, or challenging. Reaching out to the right person and getting your doubts resolved is a smarter approach than being stuck on it for hours.

Now, how to ask for help effectively?

  • Do your homework: Research thoroughly about the question that trots your mind until you truly feel helpless and need someone else’s help, then reach out. There is a thin line between asking for help and just seeking knowledge. Make sure you gather the knowledge yourself as much as possible before reaching out for help.
  • Formulate your question to avoid confusion: Make sure you formulate your question properly until no confusion is left.
  • Respect, humility, and professionalism: Remember you’re the one that is asking for help, and like every learner, you need to be ready to accept criticism and lend your ear to your helper. Be as attentive as you can. If you do not understand something or do not get it, then ask for more clarifications. People might be busy, so leave them your question and throw in a gentle reminder if they don’t answer within the week; make sure you don’t pester them.

Mastering asking on LinkedIn

LinkedIn is one of the best sources on the internet to find the right people for help and suggestions while learning Data Science. Reaching out to the right people in the right way is an essential skill that will maximize your chances of getting the solution you’re looking for- be it a dataset or any other challenge.

1. Identify the right people

The first step is to connect with the right people. “Right” here not only implies that the person should have the skill or the knowledge on what you want to ask, but also the person should be active on the platform to reply. So use the search bar to search for the keywords of the skill such as “Machine Learning”, “Data Scientist”, “sklearn”, etc. Once you get a list of people, have a glance at their profiles and their activity section. The more the number of recent activities, the higher are the chances that that person will reply. Moreover, use the “Send a personalized invite” feature to send a pre-connection message with a short introduction about yourself and your motive to connect with them.

2. Build a warm relationship

Once you’re connected, the next step is to send the first message and then build a warm relationship. This is crucial as it will help you build a relationship with that person for future help/collaborations as well, and not just a one-time affair. The first message should contain a short description of the help you need and why you think that person is the right one for it. Make sure you’ve done some pre-work and exactly know what to ask for.

A bad example of the first message:

Hi, I am stuck on XYZ dataset and not getting good accuracy. Please help me.

A good example of the first message:

Hi ABC, I see that you’ve been working at XYZ as an ML Engineer for 2 years and thought you’d be the right person to help me out on a dataset. I’d be grateful! Looking forward to your reply.”

3. Ask for help

After getting a reply, elaborate your problem in detail and ask for that specific thing you need help with. This message should also include what you tried so far or why you’re unable to proceed further. This gives the other person an indication that you’re genuinely stuck and need help. Do not bombard the person with all the details and the datasets on the first go. Brief them about the issue you’re facing and wait to see how they respond.

Ask them what other details they need to get a better understanding and give their suggestions. If you sense that the person is very willing to help and are asking more questions about the problem, ask if they’d want to connect on another medium that is better for them than LinkedIn. It could be Telegram/Slack or even a Zoom call. But make sure to be subtle and not pushy. Lastly, be patient and do not budge them very often. And in case you see that there is no reply from the person for a long time, then back to ground 0! Reach out to someone else.

A bad way to ask for help:

“I am unable to get an accuracy of 90% on this dataset (attached). Can you please help?”

A good way to ask for help:

“I am working on ABC dataset and have done pre-processing and some feature engineering. I trained a Random Forest model on it. However, the accuracy is not increasing beyond 80 even after tuning the hyperparameters. It’d be great if you could give any tips?”

4. Act on it & update the helper

Once you’re satisfied with the help and suggestions you got, it’s time to apply them. Do make sure that you’re clear with what they intended to say so that there are no gaps in your understanding.

Once you’ve acted upon it and have successfully found your way through the problem, it’s necessary to update the person that their suggestion helped and you are now proceeding on the problem. This will add another layer to your relationship with them and will indicate to them about your sincerity and that the time they spent on your problem wasn’t wasted.

However, there could also be a case that the suggestion they gave didn’t work. The best way is to again reach out to them and tell them how you applied the suggestion and what were the results. In short, keep the discussion going until you get what you need.

Lastly, it is important to end the conversation on a positive note so that there is scope for a conversation later. When you do this, it’s no longer a one-way help, but a genuine relationship.

A bad way to end the conversation:

“Thanks”

A good way to end the conversation:

“Thanks for your help/suggestions. It helped me and now I’m onto further challenges. Moreover, I’d be more than glad if I could be of any help to you now or anytime in the future. Cheers!

Seeking help via Emails

Emailing is another easy way to reach out to a person for help. One downside of emails is that not everyone reads all their emails regularly and replies promptly. But why let go of an opportunity without even trying?

The first email

The first email you send should contain all the information regarding what you need from the person. The subject should be concise and apt as well. The email should be well structured as below:

  1. Your background
  2. Why you think they can help you
  3. Describe the problem and ask for help
  4. End on a hopeful note

One mistake that many people commit while emailing someone is that they assume their problem is very silly and a waste of time for the helper. Many people begin their emails by apologizing, which is absolutely not required.

A bad way to start an email for help:

“Sorry to disturb you, but I need some help on…”

“Really sorry to take your time but I am stuck in an ML problem..”

A good way to start an email for help:

“Hi ABC, I found your email from XYZ and thought that you’d be the right person to reach out for help on a dataset that I’m working on…”

Consistency is key: The gold is in the follow-up

Likely, the person doesn’t reply promptly. People are busy, or might just not be active on emails all the time. The best way to get you what you want is to simply follow up. Once, twice, thrice, keep following up by just adding a line indicating that it’s a follow-up to your original email. Keep a gap of a few days in between your emails just to make sure you’re not pushing. Never think too much about what the person will think of you if you repeatedly email them. People hit jackpots just by cold emailing multiple times which sometimes even impresses the other person with their consistency.

Asking for Help on the Job

Another setting where the art of asking for help is important is when you start your new job. Let’s have a look at the most common reasons for not asking for help and how to get past them:

Fear of Looking Stupid

Overthinkers would know this situation very well — you are about to ask for help but the fear of looking stupid keeps running in your head. “What if my question is a stupid question?”

Often, people welcome your questions and are happy to share. People like to be asked for help and it could help you build relationships at work.

Sometimes we encounter people who can be discouraging or unwilling to help — it happens, don’t take things too personally. It reflects badly upon them and not you.

Wanting to resolve problems yourself

Being willing to tackle problems on your own is an excellent trait. You learn a lot more when you think for yourself and try things out. But if the problem remains too difficult to tackle, knowing when to give up and reach out for help is also an important trait. Often, a more experienced colleague would be able to solve your issue easily, saving you a whole lot of time spent experimenting. Also, there can be a lot to learn from observing how someone more experienced debugs or resolves any issue — these are valuable lessons that you’ll likely not find in school or online!

Don’t know what you don’t know

Sometimes we find ourselves in situations where we are completely lost and don’t know where to start asking questions, and just give up completely, thinking “I’ll understand this once I spend more time on it”. In this case, we want to get an opinion from another data scientist. Some important information to share while asking for help include:

  • The problem you are working on, and what is the objective?
  • What is the context of the dataset?
  • The success metrics?
  • What methods and models have you already tried?
  • What issues are you facing?

Conclusion

Phew! We covered a lot of ground here. But this all will be beneficial in your career — be it during upskilling, job hunt, or in just any stage of your life. Once you master the art of reaching out to people freely and are able to put your point ahead clearly, problems will slowly start to fade away. Not only will it help you achieve your goal quicker, but it will also put you much ahead of the competition, just as we saw in the Danny-Pat case. So keep this article as the guide and come back to it whenever you feel unsure how to proceed.

Ready to test your skills?

If you’re interested in collaborating, apply to join an Omdena project at: https://www.omdena.com/projects

Related Articles

media card
Building Soft Skills to Stand Out in your Job – By Computational Linguist, PhD, Anna Koroleva
media card
Tips for Data Science Enthusiasts: From an Engineering Degree to a Senior Data Scientist
media card
How to Get Involved in Data Science for Social Good (DSSG)