Author Topic: Spatial distribution on Kuala Lumpur map  (Read 14088 times)

hhyoong

  • phpBB Newbie
  • *
  • Posts: 18
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #15 on: January 21, 2011, 06:42:55 PM »
@Noordin

The water sources are actually from pipe water all over KL.

Do you have any idea which model would best fit the distribution?

Can you tell me what should I put for the Z-value field? Is it my impurities level? And what does the power value means? Assuming that is in IDW mode.

MyGIS

  • http://www.facebook.com/groups/mygis
  • Administrator
  • phpBB Newbie
  • *****
  • Posts: 238
  • http://www.facebook.com/groups/mygis
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #16 on: January 21, 2011, 11:54:02 PM »
Hi,
Do you mean that you go to places like houses offices ...etc , take water samples from taps than thoughtsyze that water?

Thanks.
« Last Edit: January 22, 2011, 01:47:52 AM by MyGIS »

hhyoong

  • phpBB Newbie
  • *
  • Posts: 18
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #17 on: January 22, 2011, 02:24:41 AM »
Hey.

Mostly from restaurants and petrol stations.

Trio of us had done the lab thoughtsysis for 13 elements/parameters as required by our supervisor. What we need to do now is to do the spatial distribution for each single element based on the result into KL map. Therefore, we have to produce 13 KL map with spatial distribution on it.

Pictures below are the one we managed to produce so far based on your guidance.


Original map with the coordinates


This is the one with IDW model.

I'm not sure whether this is the right way. Any idea how to fit the contour into the map instead of rectangular shape?

Please guide. Thank You
« Last Edit: January 22, 2011, 02:35:21 AM by hhyoong »

MyGIS

  • http://www.facebook.com/groups/mygis
  • Administrator
  • phpBB Newbie
  • *****
  • Posts: 238
  • http://www.facebook.com/groups/mygis
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #18 on: January 22, 2011, 01:17:55 PM »
Hi,

In my opinion the way you are representing your results is inadequate. To show the contrast, you should simply use some qualitative method to symbolize your points. I would assume that you are using ArcGIS Desktop for your work. Go to the layer property , symbology and choose Graduated Symbol or Proportional Symbol method under the "Quantity" group of symbology.

In your method above you are assuming that the phenomena is spatially related somehow for all the restaurants  neighboring the ones you have sampled which in my point of view may not be the case.

Perhaps Noordin may agree.

Thanks.

hhyoong

  • phpBB Newbie
  • *
  • Posts: 18
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #19 on: January 22, 2011, 04:15:36 PM »
Thanks for the reply.


Graduate Color


Graduated Symbol


Proportional Symbol

To me, all three above are not spatial thoughtsysis.

However, our main objective is bout doing spatial thoughtsysis on a KL map and produce color contours to indicate the level of parameters as requested by our supervisor. So I think we have to assume that they are spatially related. Let's ignore whether it is adequate or accurate.

Can I know how to fit my color contours into KL map? I don't want it to be out of the boundary of KL.

Please guide. Thanks once again
« Last Edit: January 22, 2011, 04:19:08 PM by hhyoong »

Noordin

  • phpBB Newbie
  • *
  • Posts: 24
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #20 on: January 24, 2011, 08:05:02 PM »
You need to have the boundary of KL as a layer. Then used this layer to mask the interpolation or IDW that you carried out.

hhyoong

  • phpBB Newbie
  • *
  • Posts: 18
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #21 on: January 25, 2011, 05:06:27 AM »
Hi Noordin!

Can you please describe it more detail as I don't really understand.

Can I use the map that i pasted in previous replies?

Is that the boundary line (chocolate colour surrounding Kuala Lumpur) that you mentioned?

If is not how should I draw it?

Please guide. Thanks!
« Last Edit: January 25, 2011, 05:10:30 AM by hhyoong »

Noordin

  • phpBB Newbie
  • *
  • Posts: 24
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #22 on: January 26, 2011, 05:08:48 PM »
Creating a masking layer

1. Click the Show/Hide ArcToolbox button  on the Standard toolbar to show ArcToolbox.
2. Expand the Cartography tools.
3. Expand the Masking toolset.
4. Right-click the Feature Outline Masks tool and click Open.
5. Click the Input Layer drop-down arrow and click a layer or click the Browse button  to choose a layer file.
6. Type the name and location for the Output Feature Class or click the Browse button  to create it.

Using a masking layer
1. Click the Add Data button  to add the masking layer to the data frame.
2. Right-click the data frame in the table of contents and click Advanced Drawing Options.
3. Check Draw using masking options specified below.
4. Click the layer you want to use as a mask in the Masking Layers list
5. Check the layers you want to have affected by the mask in the Masked Layers list.
6. Click OK.

hhyoong

  • phpBB Newbie
  • *
  • Posts: 18
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #23 on: January 27, 2011, 01:34:04 AM »
Thx for the reply. I will get back to you soon.

hhyoong

  • phpBB Newbie
  • *
  • Posts: 18
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #24 on: January 28, 2011, 03:32:47 AM »
Thanks alot Noordin. It really works with the masking feature.



However, I still have some questions. Can you tell me what are the power and search radius type & settings?

And based on the attached image, the contour din cover the whole boundary. Is it due to lack of coordinates at that particular area (without contour)?

How do I cover the whole masking layer?

Sorry for asking so many questions.

Noordin

  • phpBB Newbie
  • *
  • Posts: 24
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #25 on: February 01, 2011, 12:43:44 AM »
For the output, the issue is simply because your output depends on your extreme limit of your sampling point. Thus your data need to be outside the boundary for you to mask the area.

The output value for a cell using IDW is limited to the range of the values used to interpolate. Because IDW is a weighted distance average, the average cannot be greater than the highest or less than the lowest input. Since the influence of an input point on an interpolated value is distance related, IDW is not "ridge preserving".

MyGIS

  • http://www.facebook.com/groups/mygis
  • Administrator
  • phpBB Newbie
  • *****
  • Posts: 238
  • http://www.facebook.com/groups/mygis
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #26 on: February 07, 2011, 04:52:58 PM »
Hi,

Please note that I find this method used erratic somehow. The phenomenon is based on data picked from linear networks. Furthermore, the data must have a temporal stamp into it. Unless if you have taken all the samples during the same hour or even day than I would probably accept those data as being from the same group of individuals (statistically speaking).

Using your methodology, you are giving the impression that in between two streets spaced by 5 kilometers there is a water contamination relationship.  ??? . I believe this is incorrect unless proven otherwise by taking intermediary samples.

If you want to use an IDW model. The sampling must be undertaken following a systematic grid for initial sampling, than using a smaller grid to conclude your findings.


I am sure this would had been more appropriate to decide on the cause of the water contamination. Is it the source? or perhaps the pipe which is absorbing (from a hole) the contaminants.

I hope this would give you a different perceptive on your research.

Good luck.
Lyes
  
« Last Edit: February 08, 2011, 02:03:41 AM by MyGIS »

hhyoong

  • phpBB Newbie
  • *
  • Posts: 18
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #27 on: February 10, 2011, 03:45:51 AM »
@ Noordin

Thanks and sorry for late reply as I was away during CNY.

If the data is outside of the boundary then it would not be part of Kuala Lumpur. That's the problem ><

Is there any other feature/method that will produce a proper contour?

@ MyGIS

Thanks for the reply. I really appreciate it. I feel that it is inappropriate as well. I will discuss with my members and supervisor as well.

hhyoong

  • phpBB Newbie
  • *
  • Posts: 18
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #28 on: March 05, 2011, 07:00:08 AM »
Hi all!

Finally, I have discussed the spatial result with my supervisor. The problem now is that the map that I used as the masking boundary has slight difference in coordinate system compared to the gps coordinates. The map was taken from GIS official site. My supervisor afraid that it would not reflect the real locations. How do I match my GPS coordinates (Garmin) to the map taken from the official site?

Please guide. Thank You!

MyGIS

  • http://www.facebook.com/groups/mygis
  • Administrator
  • phpBB Newbie
  • *****
  • Posts: 238
  • http://www.facebook.com/groups/mygis
    • View Profile
Re: Spatial distribution on Kuala Lumpur map
« Reply #29 on: March 06, 2011, 11:15:06 AM »
Hi,

What do you think then? Are your points taken really offset?

Lyes

 

Error. Page cannot be displayed. Please contact your service provider for more details. (6)