Time+series


 * To gain achieve you must cover:**

a) ensure you ALIGN your moving mean so it is clearly between the values it needs to be. b) you do not need to ALIGN the CMM it should clearly be between the first and second values c) make sure the MM and CMM finish at the right places d) your Ind. SE is the raw data minus the CMM e) **you must START your Av. SE next to the FIRST Ind. SE** f) When working out your Av. SE use =average( and put comma's between each value)
 * ** Data **

a) You must FORMAT your graph by right clicking on the horizontal axis and selecting 'text axis' and 'on tick marks' b) label the axes and give it a title. c) make sure the 'key' for the lines (just to the right on your graph) are labelled appropriately. d) you need the equation and R squared value by right clicking on your trendline. e) put your graph on its own page by right clicking somewhere near your graph and move chart.
 * **Graph**

> > y = -0.0778x + 83.867 + Average Seasonal Effect (for Sep) > > Because Sep 2007 is the 29th quarter, we substitute x for 29 then solve. (2.2 is the Av SE for Sep) > > y= -0.0778 x 29 + 83.867 + 2.2 > > = 81.6108 +2.2 > > = 83.86108 > > which means that I predict in Sep 2007 the sales of avocados will be approx $83861 (rounded to nearest $ - because it was in 1000's)
 * 1) **Purpose Statement: where you give background info on the topic you are investigating and clearly explain what you are investigating and don't forget to include that you will predict long term and seasonal trends in weight (or whatever you are investigating).**
 * 2) Trend Statement: must include whether the data is increasing or decreasing, by how much (get this from the equation, it's the first number that is in front of x) and over what period (this depends on if you are working with months, quarters etc.). It MUST be in context so use units when necessary. Example: The equation of the trend line is y = 4.9588x + 334.85. This shows us that the retail sales of recreational goods are increasing by 4.9588 million dollars every quarter.
 * 3) Seasonal Pattern Statement: Describe when there are peaks and troughs and WHY they occur; use your background info and be specific. ie: use dates (like months) not just seasons (eg: in summer).
 * 4) Prediction Statement: You must use the correct equation! Don't forget to ADD the Average seasonal effect to the equation! (See your notes and read through it carefully). __You must write it as an approximate value (Ie: Hence, I predict that the mean weight for pukekos in Dec 2012 will be approximately 985grams.__ See the example below: Prediction Answers for Avocado Sales for Sep 2007:

Using the data in 'Intro to Time Series' (below) use excel to do a 3 pt MM, then a 4 pt MM with a centred mean (use notes from your green booklet, pgs 8-11)
 * 19 March 2013 **


 * 20th March 2013**

Work from Green Booklets, pg's 8-10. You need to enter the data onto an excel spreadsheet (open one above if you can't find one)

Make sure you can show a 'trendline' (pg 10 of your green booklet for 'how to'). Go through all the steps, up to trendline for this data:
 * 21st March 2013:**



Note: If you graph does not look right, ie: the axes are wrong; you change it by Right clinking somewhere on the graph, then click on 'select data'. A table will appear, you want the vertical (labelled legend) axis to have the data column B and the horizontal axis to have the dates (column A). Do this by selecting 'Edit' then select and drag the column with the dates.

We must be able to comment, in context, on your graph, describing the trend (ensure you use increase or decrease too). Look through this exemplar:
 * 22 March 2013**

Here is an example of a comment from the above exemplar: The equation of the trend line is y = 4.9588x + 334.85. This shows us that the retail sales of recreational goods are increasing by 4.9588 million dollars every quarter.
 * ** Trend of the time series has been described in context. ** ||

Work on Seasonal Effect, pg 13 of Green Booklet. Type the data onto an excel spreadsheet (just like it is on pg 13). Now find the 'Average Seasonal Effect' for each quarter. Also, describe the overall trend.
 * This year you also need to describe 'seasonal patterns' in context, state why there are peaks (justify your answers). Additionally, we need to make a forecast **
 * 26 March **

Now Do the same for the following data (use a 4 pt moving mean). For 'History of the Internet', you do not need the second column 'month number' delete or just ignore it. For 'Kiwifruit exports' you will need to switch the axes (as described above in 21 March lesson).



Use the data on Pg 13 of your green booklet (copy the first two columns on an excel spreadsheet). Then work through to finding the Seasonal Effect, then the Average Seasonal effect for each quarter.
 * 27 March**

Now graph it, find the trendline and describe it (see practice comments above in March 22 lesson).

When you have finished that do the same for the following data Remember: If the data is quarterly, we use a 4 point moving mean, if its a cycle over months of a year we use a 12 point moving mean. Also, if using an even MM we have to find the Centred Mean too.

Use this data set to find the Average Seasonal Effects (more dates to use) (Not all answers there, also add labels and title for the graph)

Pg 14 and 15 of HW Book

We'll be working on Predictions and forecasts next week.

Continue what we were doing last week (Seasonal Effect and Average Seasonal Effect; as well as describing the trend).Finish work from 27th March if you haven't. Then work on the data below **(Note: you may need to switch the axes (you'll know cause the graph will look strange); see 21 March lesson to see how).**
 * 3rd April**



We will be going onto predictions and forecasts. When you are ready read pages 19 and 20 of your HW book, read it carefully then read it again. Try the questions that follow.

Now that you are confident with Seasonal Effects and Average Seasonal Effects we are moving onto __Prediction and Forecasts.__ Read pages 19 and 20 of your HW book, read it carefully then read it again (The questions that follow is to be completed for HW). Now in your Green book, read pg14. Note our starting value is NOT going to be from zero but from number 1 (this is because our excel program starts from 1, not 0).
 * 4th April**


 * Making Predictions Practice. **


 * 1) Insert column on left
 * 2) Number from 1
 * 3) Now continue seasons until the date required for prediction: be careful, make sure you highlight all the seasons in the pattern before you drag down.
 * 4) This will give you the number of intervals required to predict ahead from the initial value.
 * 5) Substitute this value into equation and then add the seasonal value.
 * 6) This gives you your prediction

e.g. Predict 26 seasons from initial value for the equation y = 0.7x + 4.5 with a seasonal effect of -2.8 for June

y = 0.7x **26** + 4.5 - 2.8

//This is the **number of intervals** This is the **seasonal effect for June**//

= 19.9


 * Practice Predictions **
 * 1. Drivers License **
 * Practice calculating seasonal effects (that always means finding the Average Seasonal Effects too) use “Drivers License” without seasonal effects.[[file:AS 3.1 Drivers License without Seas effects.xls]]
 * Then, make a prediction for March 2006

Prediction Answers for Avocado Sales:
 * 2. Avocado Sales **
 * This time you must first find the moving means use “Avocado Sales” data.[[file:Avocado Sales Data.xls]]
 * Now, calculate seasonal effects (Average Seasonal Effects)
 * Now, make a prediction for Sept 2007.
 * Check this data here to see if your calculations were right.[[file:Avocado Sales with Seas Effects.xls]]

y = -0.0778x + 83.867 + Average Seasonal Effect

Because Sep 2007 is the 29th quarter, we substitute x for 29 then solve.

y= -0.0778 x 29 + 83.867 + 2.2

= 81.6108 +2.2

= 83.86108

which means that I predict in Sep 2007 the sales of avocados will be approx $83861 (rounded to nearest $)

Using the Drivers license data below. Write up your investigation, from start to finish (practise assessment). You are investigating the % of drivers not carrying a license and making a predication for March 2006. Save it, so I can check it on Tues. HW is also due tomorrow (Tues, 9th April).
 * 8th and 9th April 2013**

READ THIS:
 * A short cut to finding Average Seasonal Effects (similar to when we find the moving mean), is to type =average( then click on the Ind. seasonal effect cells for June (as an example) using a comma in between. Eg: it could look something like this =average(F7,F19,F31,F43)**
 * This saves you a LOT of time cause you can simply drag that first cell you made to find all the other averages (like we do for the moving mean).**



Finish your writing comments for the Drivers License data (as well as the graph and Ave Seasonal Effects); make sure you include: your purpose statement, overall trend, features of your seasonal patterns (raw data) and your forecast (predication for March 2006). **See the exemplars for some ideas.**
 * 10th April 2013**
 * Our next practice assessment (max of two lessons to work on it) is on whale sightings: background info from handout given out at lesson.**
 * Create an appropriate purpose statement (using the background knowledge you know as well). Remember to comment on features of your seasonal patterns (raw data), overall trend and your forecast is to make a prediction for whale sighting in April 09.**


 * Check your answers here, don't worry about any strange looking graphs (those are for excellence):**

**11 April (HW: finish this investigation - print it or email it to me nathans@freyberg.ac.nz prior to Tue, 16 Apr)**

 * When you have finished writing up your comments (and graph) for 'Drivers License data' Go onto the 'Electronic Card Transactions'. For this investigation check that the steps below are completed (use the exemplars to help you):**


 * 1) Develop a purpose for the investigation.
 * 2) Select one of the variables to investigate.
 * 3) Display your data in an appropriate way.
 * 4) Identify features in the data and relate them to the context (Eg: comment on any unusual features as well & try to explain why they might have occurred).
 * 5) Develop an appropriate model based on the trend and seasonal considerations.
 * 6) Use your model to make a forecast for Dec 09 - how confident are you about your forecast?
 * 7) Write a conclusion.



You can check the answers to 'Electronic Card Transactions here (don't worry about strange looking graphs, those are for excellence:

We will be doing the same investigation for sunglasses data, as we did for electronic card transactions - look at the steps above. The forecast date we will be predicting for sunglasses is for March 06. You can check for answers in 'Shared Documents' under 13MS, then go to 'more answers' and look for 'Sunglasses graphs'.
 * 12 April (don't forget you can get your info for our assessment from E1 at lunchtime on Monday, 15 April)**

Answers for sunglasses here, don't worry about strange looking graphs, those are for excellence:

Ensure you finish the 'sunglasses' assessment above and then work on I've selected this practice as there is an answer exemplar below that you can look at to see what 'comments' and written work you are expected to do for your assessment for Achieve level. Hence, open the file below and look at the purpose statement to get an idea of the background information you need to gather for the upcoming assessment. You need to calculate a predication (or forecast) for March 2001.
 * 16 April**
 * Finish this for your HW as there is one last practice tomorrow.**




 * 17 April**
 * One final practice:** 'visitor arrivals to NZ'. See the data below. You can find the data AND background info in Shared Docs. There are also answers, for you to check, in 'more answers' folder.

Again work through the following steps (use your exemplars, to help you):


 * 1) Develop a purpose for the investigation.
 * 2) Select one of the variables to investigate.
 * 3) Display your data in an appropriate way.
 * 4) Identify features in the data and relate them to the context (Eg: comment on any unusual features as well & try to explain why they might have occurred).
 * 5) Develop an appropriate model based on the trend and seasonal considerations.
 * 6) Use your model to make a forecast for July 2011
 * 7) Write a conclusion





<span style="font-family: 'Times New Roman',serif; line-height: 0px; overflow: hidden;">NZ visitors arrivals answers here:

Practice Assessments:





See template page for an example of a conclusion/summary.

Here are answers for you to check



use a 12pt moving mean for this one