- Losses — Actually, nothing special here. Cross entropy
For some reason, Jack in the Crack’s chicken sandwich looks — and feels — like fried tofu. Even if its “late-night menu” clientele is sometimes too smacked to notice the difference, we gotta take demerits for the perfect pre-fab roundness of the patty. This isn’t a school lunch, y’all.
Recently, I’ve begun to notice that imposter syndrome is not only for people that aren’t quote-unquote “qualified” to become Data Science because I’ve met people that have been studying math, computer science, stats, etc that are suffering from the same feat. In fact, I’ve spoken to people that work as data scientist that say they feel like they are faking it.
With little education and lack of a network for guidance, I staggered my way into Data Science. Although I still have plenty to learn, I am at a point now where I can look back at where I was and be proud to say I’ve accomplished something I said that I will do, even though the odds were stacked against me.
Ensembles — After training more than 75(!) models, most of them achieving AUC better than 0.9 (and the best reached 0.98), we had to improve their combined result. We assumed that after training on different representations, augmentations, models and splits, their errors would be uncorrelated enough to make a better prediction all in all. We tried xgboost and other famous algorithms — the best one was eventually xgboost (Surprise! Surprise!).
Data balancing — this was a BIG issue for us. We realized that this was a key factor in this challenge. First, because the given data was unbalanced, and second, because we saw that many samples of the data are taken from the same track. Hence, they were very similar to each other. Putting them in the training set and in the valid set is a big no-no. We created a split that dodge that bullet by carefully picking segments for the validation and training set that do not share a common track ID.
Another way to overcome imposter syndrome is to convince yourself that you’ve got the grit to figure things out. I personally believe the best way to do this is by doing your own end-to-end project.
Data Science is a very new field and is not very well defined at the moment. There are discrepancies from company to company about what a Data Scientist is, hence many feel as though the path they’ve taken to end up where they are has been non-linear. You aren’t the first and will not be the last, but you should definitely stand on the shoulders of one that has gone before you.
A funny trend I’ve noticed from the podcast I listen to (Learn more about these in How I stay up to date with Data Science) is that many people do not feel as though they’ve taken a definitive path into Data Science and that they’ve almost stumbled their way here. This makes absolute sense!
In order to use your difference to your advantage, you have to be able to package your experience in a way that sits well with the hiring manager and company. You may not have the academics or experience to back you, but something undeniable you have is grit, but you must package this grit into something tangible.
If you are currently in a field where there is no direct translation into Data Science, you are very unique and you can use this to your advantage. This is not to say that people who are in related fields (or academia) are not unique, instead what I mean is that you are unique in a sense that you have not followed a direct path, so you’d have to show a lot of character to compete in the job market with someone who has.
The Kang’s decision to switch up its poorly received TenderCrisp formula in 2017 paid dividends. A smaller chicken breast and a new marinade keep the meat fresh and juicy. And who doesn’t enjoy juicy meats?
The key here is to package your difference, there is no point stating on your CV “I have a lot of grit” or “I am willing to learn” because these are generic things that I personally think should be expected of someone in a technical field.Not only is the fast food titan coming through with the paltriest lettuce game in the industry, but its very insistence on selling a sandwich when the McNugget already exists leads us to question its marketing acumen in the first place. Do people really need mayo that bad?Transitioning careers is much more difficult and scary than changing employers. However, I believe these precursors are what helped me in my transition phase and can be of great benefit for anyone trying to transition into Data Science.
Think about it, if you saw someone that built a car from scratch and they have not followed a linear path into car manufacturing, but the car is as good, or if not better, than someone who did, wouldn’t you be intrigued by how they did it? Wouldn’t you be happy to have this person on your team, knowing that with limited resources and untraditional education/experience, they are resilient enough to seek solutions? Exactly.
It’s impossible to corroborate its claim of being the first in the game to throw a rooster on a roll, but these problematic faves have been the godfather since MLK was alive. These days, the Spicy Deluxe in particular is the crown jewel. That combo of butter bun, crunchy lettuce, and Chick-fil-A Sauce combo (basically a mashup of mayonnaise, barbecue sauce, mustard, and honey) is a go-to. Just don’t fuck with the Sabbath.
Real talk, if you want to make the top five, all you need is pickles. Not only does its briny hit even out the (shudder) creamy tang of mayo, but they help cushion your gums when that preternaturally crunchy breading inevitably shatters. Your periodontist will thank you.
However, I’ve never directly answered the question of “How do I successfully transition into Data Science?” which is exactly what you will learn from this post (from my experience).
The person does not have to have transferred from the exact same field as you. I am still yet to see another ex-post man turned Data Scientist, but I wasn’t looking for that. I was just looking for someone that has made the transition.
Fuck the “chicken sandwich wars” — while Chick-fil-A and Popeyes were trading shots, Wendy’s minded its ginger-head-ass business and kept its chicken sandwich as bankable as it’s been from the jump. Like most of its menu, the chicken sandwich is just a classic mayo-lettuce-tomato situation, but Wendys plays the hits well. And we have to give it up for an above-average quality of chicken to boot. It really makes a difference. Try its new one if you’re curious, but don’t feel ashamed about sticking with the hits.
Impost syndrome can have you stagnant, hopeless and questioning yourself. However, part of transitioning into Data Science is about recognizing the qualities and experiences that make you qualified for the role you desire. For me, being the captain of every single football team I’ve played for is what gives me the confidence to be vocal, and as a postman I developed the ability to be extremely thorough and pay close attention to detail which helps me when I am trying to code an algorithm (such as in the Algorithms From Scratch Series) or when I am carrying out error analysis and more.It’s been 2 years since I last delivered mail. Quite frankly, I am not missing it, in fact, the thought of me being a mailman seems absurd to me. But it definitely happened, I was first working for Royal Mail which is Britains postal service and courier, before switching to the mailroom at the City of London’s Guildhall.You may be glad to know that you aren’t the first person in the world transitioning from another field into Data Science. Find someone else that has done it before and listen to how they did it — This is not a personal plug.
- Date of publication:
- Sun, 11/22/2020 - 09:20
Click on the link - it will be copied to clipboard