| Author | Message | ||
Anthamidhya Junior Artist Username: Anthamidhya Post Number: 474 Registered: 02-2017 Posted From: 70.119.103.67 Rating: N/A |
Fraud alerts, recognizing money laundering acts, services side credit modeling ilaa, optimal resource allocation, customer service lo regulatory compliance etc etc saala cheyyochu | ||
Saidabad Junior Artist Username: Saidabad Post Number: 358 Registered: 07-2017 Posted From: 173.36.240.168 Rating: N/A |
2 things... 1) preassumed concepts ni prove cheyataniki intha srama avasaram ledu, maname edo cookup cheyonchu, thats not even a consideration 2)business acumen chala stron vundali besides business chala ardham kavali nenu chebutondi ade bhayya...very few practical use cases vunnayyi ani, prati daniki vadalemu, vadina earth shattering thought provoking results ravu
hmm so this is happening, I was looking for such insight | ||
Gochi Side Hero Username: Gochi Post Number: 3959 Registered: 12-2006 Posted From: 64.121.67.210 Rating: N/A |
vachina answer manaki nacchakapothe we cannot say its not success...probably there is no pattern in the data at all...intha money spend chesaam kaabatti show me some results ante mana desi companies ni pettukondi, meeku kaavalsina result choopisthaai (result ni mundey coding chesi expected result oche laaga)...Data Science/ Machine Learning is all about identifying a pattern or behavior...we cannot forcefully bring out a behavior..nor we can go with a presumed conclusion and ask data analysis to support it... | ||
Redbull Side Hero Username: Redbull Post Number: 5879 Registered: 03-2009 Posted From: 81.152.59.168 Rating: N/A |
spot on..doubt emi ledhu..already most of banking world lo ee data science ani last 2-3 years sevulu moosi saava gotti ipudu madatha peduthunnaru.. end users..emi saadistundhi vayya nee solution ani pedhavi virustunnaru | ||
Redbull Side Hero Username: Redbull Post Number: 5878 Registered: 03-2009 Posted From: 81.152.59.168 Rating: N/A |
so called data science and moderen AI etc..have just 10-15% use cases. ee banking lo data science is pedha boothu..what the heck u want to do with customer transaction data. from customer point of view all i need is good saving rate on my savings and less interest on my borrowings.. amazon, social media feed is different use case..same banking vaallu guddigaa foloow ayipodham ani choostunnaru.. | ||
Janasena Junior Artist Username: Janasena Post Number: 34 Registered: 03-2017 Posted From: 170.49.113.17 Rating: N/A |
3-4 months is very low.. Normally the training / exploration/ takes 4-6 months based on the dataset and complexity. | ||
Janasena Junior Artist Username: Janasena Post Number: 33 Registered: 03-2017 Posted From: 170.49.113.17 Rating: N/A |
Azure/Google/AWS has ML learning platforms.. edx lo courses vunnayi pure data science meeda..chosuko Side na R/Python/SaS nerchukoo You need to know which algo to apply based on use case.. | ||
Vjawarrior Legend Username: Vjawarrior Post Number: 36783 Registered: 02-2014 Posted From: 192.251.84.7 Rating: N/A |
how to start learning data science...do we need to be a good p[rogrammer to become a data scietsit...asalu ela start cheyyali...future is all data science antunnaru... blockchanin padukunnatlu undi...no one talking abt it JP Rocks: atu gaa paaare murikkaalava - aa rojule veru kada annai OT's legendary vijayendra varma review lo aa kondalu konalu loyalu sarasa sallaaaaapaalu type lo aa cycle, danikunna springu seat, kaaanthini vedajalle light, sahanaaaniki maaru peru la ooosalu, moortheebhavinchina manavatvam laati methati tires thaathaa, enti maaakeee raatha | ||
Janasena Junior Artist Username: Janasena Post Number: 27 Registered: 03-2017 Posted From: 170.49.113.17 Rating: N/A |
Data Accuracy, Filtering bad data kosam kooda ML models vaduthunnam | ||
Janasena Junior Artist Username: Janasena Post Number: 26 Registered: 03-2017 Posted From: 170.49.113.17 Rating: N/A |
Data Science fail avvatam vundadu...Data scientist fail avvochu We are working with Data scientists whose models predict with 95% Accuracy, which is very critical for the current customer business running with out any hiccups Finance sector lo Fraud alerts, and many other things kooda predict chesthunnayi successful gaa... | ||
Saidabad Junior Artist Username: Saidabad Post Number: 357 Registered: 07-2017 Posted From: 73.223.85.66 Rating: N/A |
Can you give examples of usecases that were successful and not as well, thanks | ||
Scout75 Junior Artist Username: Scout75 Post Number: 445 Registered: 08-2018 Posted From: 166.170.0.118 Rating: N/A |
from my point of view I would try to apply to atleast a few accounts and check if your data points match before and after a DS . If it didn’t may be you can analyze your missing business case . I’m might be wrong though since I don’t know the problem or the solution provided | ||
Saidabad Junior Artist Username: Saidabad Post Number: 356 Registered: 07-2017 Posted From: 73.223.85.66 Rating: N/A |
Exactly imagine my feelings when they were presenting their solution, 6000 products lo Nenu 200 skus isthe 2 solutions anta | ||
Saidabad Junior Artist Username: Saidabad Post Number: 355 Registered: 07-2017 Posted From: 73.223.85.66 Rating: N/A |
Datascience meeda I am very skeptical 🤨, yet to see any real value This thread is to learn about successes and real business results people might have seen in their projects | ||
Vjawarrior Legend Username: Vjawarrior Post Number: 36774 Registered: 02-2014 Posted From: 68.53.21.199 Rating: N/A |
mav...data science nerchukunnavaa? hyd lo IIT candis have an excellent institute for datascieence...many people learnt there and got some jobs here...later they are moving to amazon JP Rocks: atu gaa paaare murikkaalava - aa rojule veru kada annai OT's legendary vijayendra varma review lo aa kondalu konalu loyalu sarasa sallaaaaapaalu type lo aa cycle, danikunna springu seat, kaaanthini vedajalle light, sahanaaaniki maaru peru la ooosalu, moortheebhavinchina manavatvam laati methati tires thaathaa, enti maaakeee raatha | ||
Saidabad Junior Artist Username: Saidabad Post Number: 354 Registered: 07-2017 Posted From: 73.223.85.66 Rating: N/A |
One of the usecase - we want to figure out reorder points for our products Requirements are clear, data is there but solution apply chesthe results lo not much improvement. Ante memu Maa tools lo calculate chesina reorder point ki data science vallu proper chesina daniki not much dufference. So we are not seeing the value add, data Science looks good theoretically but not so succeduti when applied to real business problems ani doubt. | ||
Maverick Legend Username: Maverick Post Number: 72558 Registered: 01-2008 Posted From: 73.239.22.111 Rating: N/A |
 There's a cacophony in the truth, A melody in lies and it accompanies one on every journey, From the lows to the highs | ||
Vjawarrior Legend Username: Vjawarrior Post Number: 36772 Registered: 02-2014 Posted From: 68.53.21.199 Rating: N/A |
saidabad....amazon is hiring data scientists ..160k with 50k sign on...2-3 years exp in Data science is enuf JP Rocks: atu gaa paaare murikkaalava - aa rojule veru kada annai OT's legendary vijayendra varma review lo aa kondalu konalu loyalu sarasa sallaaaaapaalu type lo aa cycle, danikunna springu seat, kaaanthini vedajalle light, sahanaaaniki maaru peru la ooosalu, moortheebhavinchina manavatvam laati methati tires thaathaa, enti maaakeee raatha | ||
Saidabad Junior Artist Username: Saidabad Post Number: 353 Registered: 07-2017 Posted From: 73.223.85.66 Rating: N/A |
Yes!! Let’s say solution was to identify reorder point for high revenue products, either the solution is too broad and did not consider the variability associated with each product so can’t be really applied . As you said 4/10 success rate kooda choodaledu. First hire chesindi mgmt consultant ne | ||
Biggboss Side Hero Username: Biggboss Post Number: 7925 Registered: 08-2017 Posted From: 173.247.200.98 Rating: N/A |
I was not trying to highlight the technology part, I am trying to say with good Data Scientist and with clear requirements and data - it's not a hard problem to solve | ||
Biggboss Side Hero Username: Biggboss Post Number: 7924 Registered: 08-2017 Posted From: 173.247.200.98 Rating: N/A |
What exactly is your use case? | ||
Saidabad Junior Artist Username: Saidabad Post Number: 352 Registered: 07-2017 Posted From: 73.223.85.66 Rating: N/A |
Not worried about the technology they are using, I am referring to the success of the proposed solution | ||
Saidabad Junior Artist Username: Saidabad Post Number: 351 Registered: 07-2017 Posted From: 73.223.85.66 Rating: N/A |
Ofcourse, millions of past selling and buying patterns, product offerings, diff GTMs and RTMs, promotions and other discounts - ivi anni analyze chesi we want to figure out reorder points for our products | ||
Scout75 Junior Artist Username: Scout75 Post Number: 444 Registered: 08-2018 Posted From: 166.170.0.118 Rating: N/A |
How do you define success . Did you implement any solution provided or do you think that’s like a far away target to your organization. What did you predict would be coming out for your assignments given to data scientists. For example I work for a big bank . Our Business has specific targets and they ask us to provide data for that business cases . We do deliver a lot of them kani more than 50% invalidate avuthayi as in every data point they find some abnormality but we can hit our targets like 4/10 atleast in my project and they are happy about it . Again like I said it depends on how you see your business case with the solution provided and how you want to apply it . When you think the solution provided to you is not applicable at your present organization level why not hire a management consultant who can think a bit outside the box or give your data scientist a specific case and ask him the things that you want to see | ||
Biggboss Side Hero Username: Biggboss Post Number: 7923 Registered: 08-2017 Posted From: 173.247.200.98 Rating: N/A |
Are you doing SKU categorization? I am currently working on Product Categorization - our DS engineer was able to build a quick solution in a week using Tensor Flow, AWS Rekognition and get decent results  | ||
Vishvak Hero Username: Vishvak Post Number: 11498 Registered: 01-2007 Posted From: 104.129.200.116 Rating: N/A |
Is that really a data science problem? Vi veri universum vivus vici | ||
Saidabad Junior Artist Username: Saidabad Post Number: 350 Registered: 07-2017 Posted From: 73.223.85.66 Rating: N/A |
Folks Looking for feedback on successes you have seen on data science projects Initially we outsourced the effort, it took them good 4 months to understand our business case, nuances of various driving factors. When they delivered, I felt solution proposed was several feet high. 200 skus were broadly categorized into 2 groups and solution provided. Next we hired an in-house data scientist, same story is getting repeated. He was hired to apply the above solution to prove the concept as well as work on other projects. I have turned into a data science skeptic now. Looking for real time successes. |