วีดีโอ

7:39. When I have a excel files with multiple sheets, and in each sheet there's data and I conver this input file to CSV... How is the syntax to address each sheet and select and specific colum ? HUGEE THANKS Mr. Zhijing Eu :)

thanks very very nice but pl share formuale for NPV in column M ,

NotebookLM just added the ability to add TH-cam Video links so it can summarize transcripts of long YT videos ! blog.google/technology/ai/notebooklm-audio-video-sources/

Sorry about the sound quality of the GenAI Podcast clips but if you'd like to listen to the full audio I've also uploaded them here: UK Public Procurement soundcloud.com/zhijing-eu/uk-public-procurement?si=7566b0328b434148a52290bc3bd6d424& Wittgenstein soundcloud.com/zhijing-eu/wittgenstein?si=7566b0328b434148a52290bc3bd6d424&

Suppose I am calculating the uncertainty in a quality H, where as H= a x b x c x d and in here, b,c and d, all three have low, most likely and high values available, should I calculate low, most likely and high values of H first and then calculate uncertainty using three values of H? OR i should first calculate risk vales for b,c and and then some how use them to calculate the uncertainty in value of H?

Hi Zhi, Thank you for this detailed elaboration. If we have one of the cost component, for which we are sure that that has only one cost estimate, can we assume its lowest, highest and most likely value to be same and include it in the sum?

fake fuulllllls mvl gar bageeee

mvl fuuuuul doing risk without montecarlo simulations, he is a big fuuullllllll of the world

Global Business Services (GBS) are centralized models that deliver services to multiple business units across functions like HR, IT, Finance or Procurement to streamline operations, improve efficiency, and better serve customers. In 2023, Malaysia was ranked 3rd out of 78 other countries in Kearney’s Global Services Location Index 2023 for the country's cost-effectiveness and quality of talent pool. www.kearney.com/service/digital-analytics/gsli

At 8:05, I didn't manage to cover it in the video but rather than applying even more controls on how data is created, another way to "break the habit of quick & dirty local point fixes" is to introduce Data Quality Feedback Loops. These are quick report - respond - improve - communicate cycles that people using the data can easily apply to close the loop with people responsible for generating the data in the first place thereby making the value of good data clearer. This is covered in this HBR article where the core idea is to design products/ processes that continuously gain feedback throughout all parts of the (data) consumption stages and leverages those inputs to iteratively improve the user experience. hbr.org/2023/07/to-get-better-customer-data-build-feedback-loops-into-your-products

Greeting from Singapore~😀

On a closely related topic to Data Swamps - in software development there is something called the Big Ball Of Mud Architectural (Anti) Pattern which describes systems that are "haphazardly structured, sprawling, sloppy, duct-tape and bailing wire, spaghetti code jungle. These systems show unmistakable signs of unregulated growth, and repeated, expedient repair. Information is shared promiscuously among distant elements of the system, often to the point where nearly all the important information becomes global or duplicated. The overall structure of the system may never have been well defined. If it was, it may have eroded beyond recognition" blog.codinghorror.com/the-big-ball-of-mud-and-other-architectural-disasters/

For any AI enthusiasts - you can also check out a Generative AI written article on my Medium page that was based on this video here : zhijingeu.medium.com/learning-from-failure-embracing-mistakes-for-growth-079adbaff798 which was made using a service called Video To Blog www.videotoblog.ai/

00:07 Introduction 00:45 0.Importing Libraries 01:07 1.Creating Custom Class 01:46 2.Loading & Cleaning The Input Data 04:06 3.Instantiating A Custom Class Object 06:05 4.Aggregating Data To The Selected Time Granularity 07:40 5.Spend Over Time and Retain vs New Customer Over Time Plot 10:10 6.Cohort Analysis - Overview 12:14 6.1 Cohort Churn Analysis By Join Period 15:14 6.2 Cohort Churn Analysis By Customer Attribute 16:23 6.3 Average Retention By Time Step 16:54 6.4 Total Customers By Cumulative Transaction Count 17:40 6.5 Average Duration Between Sequential Transactions 18:40 6.6 Cohort Churn Analysis By Join Period By Max Inactivity Period 21:00 7. Recency Frequency Monetary RFM Analysis 23:11 Use Of Same Custom Class With Online Retail II Dataset 24:35 Conclusion

Here's the link to the Medium article - blog.devgenius.io/analyzing-customer-retention-via-cohort-analysis-1f381748e555

Great stuff

really good! Congrats!

Timecodes 0:00-Introduction 1:30-Dataset Background , Data Loading & Data Clean-Up 3:12-Selecting The Right Time Granularity 5:39-Exploratory Data Analysis 8:00-Train/Test Split 8:50-Recency Frequency Monetary Analysis 14:05-Simple Average Customer Lifetime Value 15:12-PyMC Marketing Python Library 17:00-Fitting A Buy Till You Die BG-NBD Model To Training Data For Purch. Freq & Drop Out Rate 23:31-Plot Of Probability Of Active/Inactive For Individual Customer 24:39-Fitting A Buy Till You Die GG Model To Training Data For Monetary Value 26:42-Evaluating The Model Performance Against Test Data Results 28:48-Forecasting Future Purchasing Behaviours & Customer Lifetime Value 32:27-Wrap Up

Part 1 which covers Buy Till You Die BTYD Modelling is available here : th-cam.com/video/SngwCEt_2MI/w-d-xo.html

Check out the GenAI powered multi-lingual version of this same video here : th-cam.com/video/wz4lbSK07Tg/w-d-xo.htmlsi=1xA1SP7MZiuH8_vu

Here's the full Gen AI Hindi translation app.heygen.com/video-translate/share/a44c1599af17409fa50bf2ea37a71192

Here's the full Gen AI Filipino translation app.heygen.com/video-translate/share/a060d5c87fa74631873f4ae1fb8b66c2

Here's the full Gen AI Polish translation app.heygen.com/video-translate/share/f7e0971855bb4fc69f93db2bfae2707f

On topic of misunderstandings : Wife texts husband on a cold winter morning: "Windows frozen, won't open." Husband texts back: "Gently pour some lukewarm water over it and then gently tap edges with hammer." Wife texts back 10 minutes later: "Computer really messed up now."

Can you share the python code related to XLRisk Monte Carlo simulation ?

If you liked this video you may also enjoy this other explainer videos on Wicked Problems Kennisland-How To Work With Wicked Problems th-cam.com/video/HrWbicvDLPw/w-d-xo.htmlsi=eUvg7T16_D-nIltW

Came for monte carlo , stayed for an open source tool, XL Risk !

Great guide, thanks a lot!

Hello Zhijing Eu, thanks for your kind effort. Im trying to practice the simulation by watching your video. I've used the same variables and the same values like yours. I've installed the XLRisk correctly. However, my resluts shows no outputs, no graphs but inputs only. Could you please guide me to solve the issue? Thanks in advance.

Hey Zhijing, firstly, this is amazing and very helpful. Thanks for the content. It made life easy. One thing I can't get my head around is how the correlation matrix works in understanding those values and attributing them to the elements. I have gone through your links posted for more examples and answers in the comments yet I can't seem to wrap my head around it. How does the risked value lets say in accommodation, understand that there is a correlation of 0.85 between its value and the value of meals. I do not see any link in those tabs. The function reads numbers from a table that do not have the elements linked. I might be super confused, but I've been scratching my head around this a lot. Hope you can help...

great Explanation

Can you help me with something? I'm getting the following error with corr matrix: "The table (referenced) doesn't have all off-diagonal values between -1 and 1". The thing is that it is, and there is nothing on githup about how to deal with it.

Hello Zhijing Eu, thanks a lot for your video. i try to use monte carlo for QMRA. Is it true if i say graph cummulative distribution describe about variability? I found out about how to make uncertainty with monte carlo, and it said that i need to second-order. But i confused about parameter input that i use in second-order. And i cant find Tornado diagram in my XLRisk. i use 2 parameter input for my output. does it affect the output of the diagram?

So cool!

Zhijing, amazing explanation. Very didatic. Many thanks. If you allow me please inform in the details of the video the link to download your file.

Hey Zhijing, can you show it again how you linked the correlation matrix with results in I4:I9. I am getting following message when I try to run simulation. "The risk functions have not been properly linked to the correlation matrix".

Goog job, this is great man ! Keep going

is it possible to download the spreadsheet to see the data there? recognising it would need the addin to operate. thanks

Hi idol. next please. fluid simulation of navier stokes😊

That was nice explanation. Could we be able to do it using python ? Thanks

it is not clear how you connect the correlation matrix to the original ranges

FYI only if you are stuck trying to replicate this and getting an error when calling GPTSimpleVectorIndex, it is because a few months after I published this, LlamaIndex renamed GPTSimpleVectorIndex to GPTVectorStoreIndex github.com/jerryjliu/llama_index/issues/1900

FYI only if you are stuck trying to replicate this and getting an error when calling GPTSimpleVectorIndex, it is because a few months after I published this, LlamaIndex renamed GPTSimpleVectorIndex to GPTVectorStoreIndex github.com/jerryjliu/llama_index/issues/1900

How to download data

Useful. Thanks

Nice.

Hello Zhijing Eu, Thank you for this great video. Only it is not clear to me how you wired the covariance calculations in the example in the spreadsheet (the table is clear, but where and how is it connected in the model?) Can you please specify what formula's you used and in what lines/cells in the spreadsheet and in the results pagina we can find this? That would be really very helpful. Thank you!