[Artifact]

Here is a simple plot of the may 2006 data – the box represents the interquartile range (i.e. contains the middle 50% of the price range). The line in the box is the median (50th percentile) – the notch around the median is a statistical confidence interval around the median (95% confidence level).

In this data, you can see that the prices are not distributed evenly – the bottom “half” of the box (25th to 50th percentile, or 2nd quartile) is much narrower than the top “half” or 3rd quartile. So in that month, 50% of the houses sold for between ~1.35 and 3.4 million. *sorry about the funny axis mark for 900K – didn’t feel like fixing it.

If we added another month to this, you could then not just compare the median, but the distribution of prices in those “quartiles” – still not perfect by any stretch, but gives you a little more information than just the median.

[img_assist|nid=3499|title= 92037 – May 2006 Sales Prices|desc=|link=node|align=left|width=466|height=496]

[Artifact]

Here is a simple plot of the may 2006 data – the box represents the interquartile range (i.e. contains the middle 50% of the price range). The line in the box is the median (50th percentile) – the notch around the median is a statistical confidence interval around the median (95% confidence level).

In this data, you can see that the prices are not distributed evenly – the bottom “half” of the box (25th to 50th percentile, or 2nd quartile) is much narrower than the top “half” or 3rd quartile. So in that month, 50% of the houses sold for between ~1.35 and 3.4 million. *sorry about the funny axis mark for 900K – didn’t feel like fixing it.

If we added another month to this, you could then not just compare the median, but the distribution of prices in those “quartiles” – still not perfect by any stretch, but gives you a little more information than just the median.

[img_assist|nid=3499|title= 92037 – May 2006 Sales Prices|desc=|link=node|align=left|width=466|height=496]

[Artifact]

I have had to deal with highly variable data in other fields – I think what might be interesting to plot in that La Jolla data set, if it were available, would be to plot the interquartile range, or even just plot each point within the interquartile range –

Not to insult anyone, but to explain what that is – basically divide the entire range of values into quarters – the median is the 50th percentile – and plot all values between the 25th and 75th percentile or just the 25th, 50th(median) and 75th. Still has many of the problems associated with median – but gives some insight into the distribution – limited as it may be.

[Artifact]

I have had to deal with highly variable data in other fields – I think what might be interesting to plot in that La Jolla data set, if it were available, would be to plot the interquartile range, or even just plot each point within the interquartile range –

Not to insult anyone, but to explain what that is – basically divide the entire range of values into quarters – the median is the 50th percentile – and plot all values between the 25th and 75th percentile or just the 25th, 50th(median) and 75th. Still has many of the problems associated with median – but gives some insight into the distribution – limited as it may be.

[Artifact]

I have been making the plots for the short sale monitor – looking at sdrealtor’s data here I am trying to think of an entertaining way to present this data as well –

sdrealtor – in addition to the additions/subtractions, can you post the total listings/actives? I suppose if I have even one weeks numbers of that I can work from there.

[Artifact]

I have been making the plots for the short sale monitor – looking at sdrealtor’s data here I am trying to think of an entertaining way to present this data as well –

sdrealtor – in addition to the additions/subtractions, can you post the total listings/actives? I suppose if I have even one weeks numbers of that I can work from there.

[Artifact]

Here are the figures – the first week in a while that the rate of increase in short sales is noticeably less. It is still increasing, just not as fast as the previous weeks, and still increasing faster than total listings, but not by much this week.

[img_assist|nid=3489|title= Total Listings and Short Sales|desc=|link=node|align=left|width=466|height=267]
[img_assist|nid=3490|title= Percent Change and Percent Short of Total|desc=|link=node|align=left|width=466|height=267]

[Artifact]

Here are the figures – the first week in a while that the rate of increase in short sales is noticeably less. It is still increasing, just not as fast as the previous weeks, and still increasing faster than total listings, but not by much this week.

[img_assist|nid=3489|title= Total Listings and Short Sales|desc=|link=node|align=left|width=466|height=267]
[img_assist|nid=3490|title= Percent Change and Percent Short of Total|desc=|link=node|align=left|width=466|height=267]

[Artifact]

sdr – I think you mixed up the short sale number with last weeks numbers- Those two numbers are the same as for the previous two weeks? The total listing number looks correct though.

[Artifact]

sdr – I think you mixed up the short sale number with last weeks numbers- Those two numbers are the same as for the previous two weeks? The total listing number looks correct though.

[Artifact]

Same story for the figures – the rate of increase was less for both total listings and short sales, but the short sales are still increasing faster than listings. The difference between the two seems to be pretty constant.

[img_assist|nid=3421|title= Total Listings and Short Sales|desc=|link=node|align=left|width=466|height=275][img_assist|nid=3422|title= Percent Change and Percent Short Sales of Total Listings|desc=|link=node|align=left|width=466|height=275]

[Artifact]

That is an interesting model they came up with. In their defense, they do admit it’s shortcomings – particularly:

“The most significant potential drawback to this or any model-based approach is that time series modeling assumes a predictable continuation of historical patterns and relationships and therefore is likely to have some difficulty producing reliable estimates at economic turning points or during periods when there are sudden changes in trend. BLS will continue researching alternative model-based techniques for the net birth/death component; it is likely to remain as the most problematic part of the estimation process.”

It is definitely missing on this account – not only does it have the addition of 49 construction firms in April, but it also calculated the addition of 26 financial activities – that latter number certainly includes a number of things other than mortgage companies – but I think it is safe to say that it is “slightly” overestimating the number of new fincancial companies.

[Artifact]

Pretty pictures. Since the large drop in both total listings and short sales in march, there has been a steady increase in both, but the percent increase is always bigger in short sales continuing the great linear increase in percent of total listing that are short sales (green line in the second figure).

Note: For fairness, I repeated the first figure using the same scale for both numbers to remove the scaling effect in the first figure.

[img_assist|nid=3355|title= Total Listings and short sales|desc=|link=node|align=left|width=466|height=308]
[img_assist|nid=3356|title= Percent change and percent total listings that are short|desc=|link=node|align=left|width=466|height=308]
[img_assist|nid=3357|title= Unscaled listing numbers|desc=|link=node|align=left|width=466|height=308]

[Artifact]

The plots:

A note on the second plot – the percent change is plotted as week-over-week. Any point above the grey line (zero change) means an increase from the previous week. The most notable thing here is that every week, the change in short sales is always a little bigger than the change in total listings, resulting in the steady (almost linear) increase in the percent of total listing that are short sales (the green line).

[img_assist|nid=3289|title= Short and Total Listings|desc=|link=node|align=left|width=466|height=292]
[img_assist|nid=3290|title= Percent change and percent short sales (of total)|desc=|link=node|align=left|width=466|height=292]

[Artifact]

And the accompanying pretty pictures. I added a gray line on the second figure to show the zero change point for reference – above that the numbers are going up, below they are decreasing.

NOTE: I finally had to expand the left hand y-axis! Remember in the first figure there is a x10 distortion between short sale data and total listings.

[img_assist|nid=3218|title= Total listing and short sales in San Diego County|desc=|link=node|align=left|width=466|height=297]
[img_assist|nid=3219|title= Percent change and percent short sales|desc=|link=node|align=left|width=466|height=297]

[Author]

Posts