topic Re: problem in creating space time cube for emerging hot spot analysis in Spatial Statistics Questions

problem in creating space time cube for emerging hot spot analysis

sprihharsh — Sun, 09 Oct 2016 20:55:54 GMT

Hi,

I am trying to analyze hot spot pattern in 10 year of count data for a bird species (2005-2015). The result of emerging hot spot analysis is exactly what I need now. But I am having problem in creating space time cube required for doing this analysis.

What I have is

10 year of count data done of specified transect each year with the date mentioned for each year. Now these counts were done for two months each year, july and august.

While creating space time cube, I want to give all of this as input. But I am having trouble in doing that.

1. Time field will be date, now this date is varying year to year so what am I supposed to do here, date of one year?

2. Time step interval, I want an analysis per year, so is the input 1 year, but that's failing the analysis

3. Time step alignment , will this be start_time if initially I am giving date for 2005 year? or how does this work?

Do you have any demonstration which can help me in doing this correctly? I will really appreciate any help in this regard.

Thanks,

Sprih

Re: problem in creating space time cube for emerging hot spot analysis

DanPatterson_Retired — Sun, 09 Oct 2016 21:22:40 GMT

sadly How Create Space Time Cube works—Help | ArcGIS for Desktop but there are some YouTube's done by others Haven't looked at them

Re: problem in creating space time cube for emerging hot spot analysis

JenoraD_Acosta — Mon, 10 Oct 2016 23:02:15 GMT

Hi there,

1. For the Time Field parameter, just point to the field in your data set that contains the date. It doesn't matter right now that the dates vary.

2. For the Time Step Interval, if you want to analyze on a yearly basis, you should enter 1 Year. Since you have 10 years worth of data, this shouldn't be failing ... Please provide a little more information about the failing or error you are receiving and I am happy to help pinpoint the issue.

3. Time Step Alignment doesn't require a date, just specify one of the options. All this parameter is doing is telling the tool where to start chunking up your data. If you choose End Time it starts at the most recent date in your dataset and goes backwards in time. Start Time will start at the first date in your dataset and move forward in time. The reason this is important is because of any temporal bias that may be present in your dataset - basically you are choosing whether you want that bias at the start or at the end of your cube. This seems like it is going to be less important in your analysis since you know each year only has 2 months worth of data - since only 2 months out of 12 have data, the temporal bias in each time step (each year) is going to be 83% - the important thing is that we know that If you had measurements that were more continuous during the year and then the last year was basically 83% empty, you may want to choose to remove that year from the analysis. If you would like each year to start at January 1st, you should choose the Reference Time option and enter January 1 and whatever year your data starts for the Reference Time. This way every time step will be January 1st through December 31st.

We just did a live training seminar on this topic (located here) and there is also a recording of our 2016 UC technical workshop (here). A tutorial as well as a free hands on web course are in the works! In the mean time, I hope the LTS and workshop are helpful. Feel free to contact me with any other questions you may have. I am happy to help.

-Jenora

Re: problem in creating space time cube for emerging hot spot analysis

sprihharsh — Tue, 11 Oct 2016 14:53:26 GMT

Hi Jenora,

Thank you so much for the reply and tutorial, videos. They do help in clearing my concepts but I am still stuck with my questions. Just to be clear with my question, lets say I have data formatted as below.

Route ID Latitude Longitude 2015counts Date2015 2014counts Date2014 2013counts Date2013 and so on

4s 43.123 45.321 10 8|5|2015 6 8|12|2014 7 7|21|2013

And I am using Space Time Pattern Mining tool in ArcMap.

Now while creating space time cube, I am having following questions:

1. What is Time field? I have different dates for different years and so different columns for each of those years. So which to point while giving time field.

2. Time step interval? Okay, I keep it 1 year, it shows me this error " a space time cube must have at least 10 time slices". In this case in time field, I gave date2015 column, time step interval 1 year and time step alignment end_time.

3. Time step alignment?I got the concept of End_time, start_time, reference_time. But somehow I am not able to make it run for 10 years, for all the 10 years.

Now I am bale to run this with following inputs:

1. time field 2005 date

2. time step 12 months

3. alignment reference_time with reference to august,2015.

Now this runs but it runs just for 2005-2006 data with reference to 2015-2014 data.

Please help me figure out what mistake I am doing, I get the concept but I am not able to execute those learning in my steps here.

Looking forward to your reply.

Thanks,

Sprih

Re: problem in creating space time cube for emerging hot spot analysis

sprihharsh — Tue, 11 Oct 2016 14:54:37 GMT

Hi Jenora,

Thank you so much for the reply and tutorial, videos. They do help in

clearing my concepts but I am still stuck with my questions. Just to be

clear with my question, lets say I have data formatted as below.

Route ID Latitude Longitude

2015counts Date2015 2014counts Date2014 2013counts Date2013

and so on

4s 43.123 45.321 10

8|5|2015 6 8|12|2014 7

7|21|2013

And I am using Space Time Pattern Mining tool in ArcMap.

Now while creating space time cube, I am having following questions:

1. What is Time field? I have different dates for different years and so

different columns for each of those years. So which to point while giving

time field.

2. Time step interval? Okay, I keep it 1 year, it shows me this error " a

space time cube must have at least 10 time slices". In this case in time

field, I gave date2015 column, time step interval 1 year and time step

alignment end_time.

3. Time step alignment?I got the concept of End_time, start_time,

reference_time. But somehow I am not able to make it run for 10 years, for

all the 10 years.

Now I am bale to run this with following inputs:

1. time field 2005 date

2. time step 12 months

3. alignment reference_time with reference to august,2015.

Now this runs but it runs just for 2005-2006 data with reference to

2015-2014 data.

Please help me figure out what mistake I am doing, I get the concept but I

am not able to execute those learning in my steps here.

Looking forward to your reply.

Thanks,

Sprih

Re: problem in creating space time cube for emerging hot spot analysis

JenoraD_Acosta — Fri, 14 Oct 2016 20:31:34 GMT

The way your data is formatted is not what the tool is expecting … The Time Field parameter is expecting you to point to one field with all of the dates in it so it can calculate the time span of your data. There is not an option to have your dates in multiple fields.
You are probably getting the “10 time slices error” because there are not enough records in 2015 to create 10 time steps? How many records do you have for 2015?

Basically, there should be 1 row in your table for each incident in your dataset.

I am happy to help figure this out. Can you email me at jdacosta@esri.com?