HomeОбразованиеRelated VideosMore From: Jalayer Academy

Excel - Time Series Forecasting - Part 1 of 3

2273 ratings | 742478 views
Part 2: http://www.youtube.com/watch?v=5C012eMSeIU&feature=youtu.be Part 3: http://www.youtube.com/watch?v=kcfiu-f88JQ&feature=youtu.be This is Part 1 of a 3 part "Time Series Forecasting in Excel" video lecture. Be sure to watch Parts 2 and 3 upon completing Part 1. The links for 2 and 3 are in the video as well as above.
Html code for embedding videos on your blog
Text Comments (188)
joe browning (15 days ago)
Awesome job, you saved me a lot of headaches! Thanks so much! You the man!
Honey Notwani (25 days ago)
Hi Sir, I am so happy to watch your videos, but i have a Q regarding the same, If i would like to forecast date wise and i am receiving very different volume like i have receive 2000 count 1 day and next day i receive only 900 count and after 5 days i receive again 1500 above count and this is not possible that if i received 2000 count in 01-jan-2018 so i will receive approximately 2000 or more then 2000, may be possible that i will receive 1200 count in 1-fab-2018 how can i forecast for next month for 1-Mar-2018
Kamona Kar (1 month ago)
Mohsen Hashemi (2 months ago)
You saved a mystery for me thankkkkkks
Civci (2 months ago)
Awesome. Thanks
PLAY MAKER (2 months ago)
Hi Jalayer, thanks for the video. It's really helpful. I have a quick question though. What would be the MA and CMA if I have 3 years data month wise? Should I be considering the average in 12 months at a time?
L VR (3 months ago)
What is the minimum amount of starting data required for this forecasting method to be accurate??
Ridwan Rahadi (3 months ago)
Hi, How using the method to forecast multiple SKU/Product in one worksheet
Isaac Hatch (4 months ago)
Is Ryan Gosling giving this tutorial?
Robert (4 months ago)
Would you show us how to create a genetic Algorithm in Excel?
Gururaj Hiroli (5 months ago)
can we have any theory pdf regarding this method?
Larry Zhou (5 months ago)
@11:00 why do you take the MVA of 4 quarters and put it in the 3rd quarter location? so you have 2 from the past, 1 current, and the 1 of the future.
Deepthy Praveen (5 months ago)
how do I do this at Item Level - month wise rather year/ qtr sales
Digital Nomad (6 months ago)
Please, very grateful for your video but .. so many people asking you what the name of this model is but still no answer.
王朝愉 (7 months ago)
I have a question after watched all your vids on this topic. How do I apply this method to a monthly data with seasonality by weeks? t1 is the first day of the month and it can be a Sunday. But t32,which is the first date of the forecast, will be another day, possible a Tuesday. Of course you can use the Sunday's data to forecast a Tuesday's performance. How to fix this?
صبر جليل خميس (7 months ago)
بحاجة بيانات لغرض البحث 😔
Berkan Ataş (8 months ago)
thx for the help. I have missed one thing that why have we started from the 3. cell when taking moving average?
Alok Sharma (8 months ago)
well explained!! ... although redundant...
Z touka (8 months ago)
man you're too cold.... you could've landed the idea in less than 5 min
BattleofYorktown77 (8 months ago)
One thing I noticed is that at least with my data, Year, Month, (Int/Sales), changing the number of periods included in the moving average from 4 months to 12 months, did not really create much impact. I only had 2 years of history so I kept the moving average to 4 months. I was 98% accurate or more in predicting actuals. Thanks for the refresher course!!! The market I deal in is fairly stable with slight volatility when new products are introduced or insurance companies make changes.I work in healthcare. I think an area this guy highlights that some folks dont catch is actually understanding how your data moves visually along with the influencers of the data.This is a critical step as it may lead you to a different regression model.
Raj Najat (8 months ago)
I disagree with Mr Ivan...... Whatever repetitions you have done not only clarified the subject more, but it also did a bit of revision. One should think about people who are totally alien to the subject , and not about those who know too much ABOUT THE SUBJECT ! I am compelled to say that that the author and commentator of this video is one of the greatest teachers EVER ! ( this is my first video for him, but I am a physician and I know how to evaluate something ! Congratulations to the Jalayer Academy and thanks for being kind to spend your time and money for the common and downtrodden man !
Rishil Lakhe (8 months ago)
Gururaj Hiroli (5 months ago)
Moving average model
Vercingetorix061983 (9 months ago)
Nice work, dude!
Human Needs (11 months ago)
If anybody wants hindi Excel classes https://youtu.be/xEhXUaJgA4o
Ridwan Rahadi (11 months ago)
Hi There, this video very good, Thank you, but how to make forecast for many SKU of Inventory such as 100 SKU
sgteamx (1 year ago)
Thanks for the video! What types of forecasting is this, and why was it chosen compared to say, exponential smoothing / ARMA / Weighted moving average?
Great video. Thank you
Sai Phani Theja (1 year ago)
May I know the theory and formulae behind this example. I mean what should I search for to get the theory
C Alex (1 year ago)
Good watch. next time just close your window and open your air conditioner pls. Noise outside.
GateCrashers (1 year ago)
Do you cover Multi-variate time series ?
GateCrashers (1 year ago)
holy fcuk he speaks slow ! JESUS, I had to speed it up to 2 just to not fall a fcuking sleep ! Good job, though !
Travis Preciado (1 year ago)
I'm curious do you have a video using the seasonality index? Its where you average the numbers from q1 q2 q3 and q4
kailash mehta (1 year ago)
Great video ! cover fundamental of forecasting . How to manage forecasting in excel , especially if products or sku is more than 5000.
Philip Park (1 year ago)
whats up with the fire truck in the back ground? big city? lol
Brandy Jackson (1 year ago)
Thank you for breaking this down. VERY helpful.
Sam Ting (1 year ago)
Hi, thanks for the great video. I notice when you present, you can also draw with red mark on the screen, how you did that?
Dilan Sankalpa (1 year ago)
You waste time by saying lot of words. any way thank u for the video
Hessam Harandi (1 year ago)
You have the knowledge and know how to teach. Great job and thanks a lot!
D. Nelson (1 year ago)
Good info but OMG, so slow. Please edit and speed up. It was 1:10 before it got to the point.
James Perry (1 year ago)
Thanks. Thats a great! and valuable tutorial.
TAYLOR GRAY (1 year ago)
Great videos for data analysis!
Darragh Murray (1 year ago)
Have you done any videos for the additive (rather than multiplicative) method?
How accurate is that for long periods? What the confidence ratio and error margin? If i take it to, for exemple, 2020, would it be useful?
Clive Buhagiar (1 year ago)
This series of 3 is fantastic! Thanks so much. It's really straightforward to follow and does exactly what I need.
Handi Rahmadiansyah (1 year ago)
thanks a lot
Young Rod Beats (1 year ago)
My forecasting teacher has nothing on you!
Inova Aria (1 year ago)
If you are doing so analysis why you are not rich ?
Leana Phillip (1 year ago)
awesome lesson! is there any way to include other variables such like economic factors like employment rates ?
Logesh Raja (1 year ago)
Hey thank you , @jalayer academy , Can you please let how did you decide to do Moving average on 4 periods
Allan Davis (1 year ago)
4 periods is one year, the assumption is that the data is cyclic with period 1 year
唐永 (1 year ago)
It is very useful, thankyou ,look forward to your next vedio
nikhil nayan (1 year ago)
Thankyou for the video. Really helped me to clear my understanding of time series forecasting.
Why am I here im in year 7
Damar Dwi Prasetyo (1 year ago)
hey,,@jalayer academy,,,,many thanks for this awesome methode forecast,,,but,,,,would you help me, how to forecast if yt is uneven database ex: sunday to saturday (7)...??,,,
Lorinda Forrest (1 year ago)
Good info, but you only need to explain a concept one time.
Aweng Majok (1 year ago)
Great job dude! How can we identify whether we shall follow a multiplicative or additive decomposition?
Liam Brochu (1 year ago)
Which forecasting technique would this be? Would it just be considered moving average?
Rugaju Solomon (2 years ago)
jpaokx (2 years ago)
Hi. Would it be correct if you take MA(2) instead of MA(4)? The curve in this example is obvious and I see why you took for a 4-month period. But, in real word, it is not always visible. What would be your approach for monthly predictions? Earlier, you mentioned that you'll be prepping a video with monthly data. Do you have it available? Thanks mate
Wajid Raza (2 years ago)
Hi @jalayer Academy do you have any tutorials on ARMA model?
Wajid Raza (2 years ago)
Can any one refer a tutorial on ARMA model
Roham Pourmehr (2 years ago)
Is someone playing GTA in the background?
Wesley Pollard (2 years ago)
so how do account for the human element and market conditions?
Amit Chakraborty (2 years ago)
Thank you very much. this is exactly the solution to the problem i was facing. Thanks for the step by step instructions.
4gany (2 years ago)
Hi whats the name of this forecasting method ?
Thanks a lot. Exactly what I was looking for. The only question is if my season is a week, and data is daily, does it mean MA is for 7 periods?
Martin Novak (5 months ago)
Of course :) or you could make a weekly average and then repeat for future weeks to have months... Just more time
Darragh Murray (1 year ago)
I would assume so yes.
Van Nguyen (1 year ago)
And in your case, I wonder whether the first value of "MA" will be in the 4 row, instead of 3 row as the example. Besides, it is week cycle, 7 days in a week so it is odd number, and we do not need to generate CMA, do we?
Jeffrey Gong (1 year ago)
+Jalayer Academy I would also like to know the answer to this question
Kurt Vella St John (2 years ago)
Good resource but it drags just a bit too much
Eric Abadie (2 years ago)
Thank you for the video. How do I deal with multiple seasonality? I have daily information and each week there is a predictable day of week cycle i.e Tuesday and Monday's are low and Fridays are high. but in addition to this, there is another seasonality where the 3rd week of the month is usually lower than the other weeks of the month. And of course, there is monthly seasonality where the peak month of the year is December followed by a drop in January. How do I include 3 seasonality components using this methodology?
Umaprasad Sutar (2 years ago)
Hi Jalayer, Hope you are doing fine. I wanted to check with you, would it be possible for you to upload Time Series, Regression and Machine Learning models with examples from Human Resources context in excel. Any help with the same is highly appreciated. Thanks for your help in advance.
Jose Angel basurto (2 years ago)
hi, my question is, if I have 4 year information, regarding the first year has 35 periods, second and third 52 and the last just18, how many periods should I have to include for an accurate a moving average?
Narendra Rana (2 years ago)
Great set of videos on time series. Question : Are all time series based on a multiplicative model? If no, then how did you determine that a multiplicative model would fit this data. Also, why did we not try to isolate the irregularity, why just stop at trend and seasonality. Final question, why can't we use the cma value as trend, does that not indicate a trend? thank you once again
val meero (2 years ago)
you're a life saver
P Ncube (2 years ago)
May you also highlight on how you will use excel to outline the general hyperbolic function followed by the prediction or the forecast line. (e.g it can be Yt = aXt^2 -bXt + c.
Mohammad Saber (2 years ago)
Great explanation. Thanks.
Adam Nemčík (2 years ago)
Hi Jalayer, nice tutorial. Did you ever try multivariate forecast ? Im Working in Diploma Thesis with long term forecast where usuall programs like Rapidminer, etc. have little problem with horizont more than +1. I think excell could be nice alternative for my task but i have to work with 14 variables or so. Can you give me some advice please what should be different than in this way ? Thank you
Jiggyciggy (2 years ago)
This guy sounds like the secret service agent from prison break hahahaha
Tanya Devangi (2 years ago)
Hello I just wanted to know how I should address the above if I have monthly data of last 2 yrs instead of quarterly.. plz help
Chioma Odimegwu (2 years ago)
you're awesome at breaking things down and explaining. Thank you!
Air Switch (2 years ago)
Very clear. Thank you. I cannot understand my professor at all. Also his examples are terrible and unclear.
Lumhiere (2 years ago)
hi Jalayer, First of all, i would like to commend you on this awesome video. which i find very interesting and helpful on forecasting. just a question though, can this be applied on "Daily Volumes". say, let's have sales be changed to case counts and i would need to forecast the volume of the upcoming cases that would plunge in by next week, would this be applicable? considering that case count volumes differ on a daily basis and some of the days would have vacant or completely 0 case count at all, as there would be holidays. please let me know your opinion or any alternative approach. thank you so much Greetings from the Philippines !
alexchacon777 (2 years ago)
how can you do a time series analysis with multiple variables (x). ?? thanks
Btissam Rachdi (2 years ago)
Do you have a tutorial about time series using R .. Thank you !
Alex Sabir (2 years ago)
How high were you when you made this video? lol Why are you drawing all over everything? Just point with the mouse.
Muhammad Nazri (2 years ago)
Hi Sir, I've similar inquiry as karadic; why the first value of MA(4) been placed at the 3rd quarter instead of 4th quarter?
Muhammad Nazri (2 years ago)
Hi Sir, I've similar inquiry as karadic; why the first value of MA(4) been placed at the 3rd quarter instead of 4th quarter?
Altamash khan (3 years ago)
is this trend component only use for seasonal items ?or for all items .please reply seasonal index and trend component is same or not
Colton McCormick (3 years ago)
Quarter Orders Overtime 1 53 22 2 56 23 3 52 25 4 62 31 5 47 21 6 46 22 7 48 20 8 57 29 9 43 23 10 45 12 11 46 22 12 53 26 13 38 16 14 35 17 if i have one more vary is overtime how can i do that?
linkolin kampret (3 years ago)
Could you help me, if I have 8 Quarter ? Please..
Rahul Krishnan (3 years ago)
this is just awesome!!! helped me a lot in my work!!!
soukaina mahfoud (3 years ago)
Hi! If that statistically correct if I add factors to the regression model, such as index values, and if the added variable adds significance to the model then I also use forecasted index value (in addition to the period variable shown here) to forecast my revenues?
Armando Hernandez (3 years ago)
Thanks you safe my life brother!
Gugan Babu (3 years ago)
Can u please tell me how to predict the month wise data using method. Can u please explain it by an example.....
Daniele Giorgi (3 years ago)
One question?this is holt winter method or not?thanks
Douglas Pernikoff (3 years ago)
so, i am following the time series modeling and wondering what steps to pass on if i have only annual revenues for 16 years and want to predict into year 17.  i have a chart with a trend line for the real data.  do i need to perform a linear regression model first, or what steps are appropriate to get my one year out forecast?   thank you.
Amit Singha Roy (3 years ago)
Santiago Espin Basany (3 years ago)
I learned more with this three videos than with a bunch of books. Thank you very much!
Joan Badell Giralt (3 years ago)
Very good patates amb suc!!!
Samuel Kariuki (3 years ago)
what software did you use to create this video?
Maimba Mwansompelo (3 years ago)
+Samuel Kariuki he didnt reply to my questions either.do u have his email?
Oscar Rodríguez (3 years ago)
CMA value is always the average of 2 values or it depends of others factors?
Jalayer Academy (3 years ago)
+Oscar Rodríguez yes, 2 values, Center when you have performed an even MA beforehand
Ibrahim Ozturk (3 years ago)
Great video series. I have learned and applied in bussiness. I am looking forward to see the R version of the series.
Jennifer Green (3 years ago)
What if you have yearly data and not quaterly data? It doesn't seem like St- It can be done following this method....so how do you forecast for yearly data?
Raj Malhi (2 years ago)
Is this the same as Holt Winter approach? Does this optimize Trend, Seasonality and Irregularity as the Holt Winter Approach does?
Jalayer Academy (3 years ago)
+Jennifer Green You can look for recurring patterns that last multiple years. This is often referred to as Cyclicality and dealt with in a very similar fashion as Seasonality.
Wang X (3 years ago)
any ideas on how to put confidence interval around the predictions?  Thanks

Would you like to comment?

Join YouTube for a free account, or sign in if you are already a member.