Contributor
Posts: 21

# combine data sets fro regression

Hello Everyone,

I am new to SAS. I used to use Excel for some statistical analysis, but I am in trouble with larger size data now. I am a new SAS learner. If I cannot make the question clear, or not using right SAS terms, please be nice  I tried  to create a sample data, but I ran into format issues. So, i will use the following as example.

I have a data like:  ( city names: a, b, c;  province names: x, y, z)

individual  city2010   province2010    city2011     province2011    city2012     province2012 ........

1                 a                 x                    b                   x                      c                   y

2                 a                 x                    c                   y                      d                   z

3                 c                 y                    a                   x                      a                   x

4         .......................................................................(more data)

So, it reads like: individual 1 lives in city "a" which is in province "x" in 2010. Then she/he moves to city "b" (also in province "x") in 2011, etc...

I also have a yearly province-level data with information of average income  for each province and each year.

e.g.

province "x" average income in 2011 is   90

province "y" average income in 2011 is    100

province "z" average income in 2011 is    80

The above numbers will be used for below example.

My study is about to see the relationship between the number of individuals move from PROVINCE to PROVINCE and income ratios, i.e. not city level. I would like to prepare a data like:

year   province_from  province_to      count                                  income_ratio

2011            x                       y               1         (province_to_income2011)/(province_from_ income2011)   = 100/90

2011            x                       z               0         (province_to_income2011)/(province_from_ income2011)   = 80/90

2011            y                       x               1          (province_to_income2011)/(province_from_ income2011)   = 90/100

2011            y                       z               0          (province_to_income2011)/(province_from_ income2011)   = 80/100

2012           x                        y               1           (province_to_income2011)/(province_from_ income2011)   = 100/90

2012            x                      z             ..............

.

.

So, this will be province-level for each year for each pair of provinces data.

I know this is a long story. I hope it is clear. Otherwise, please ask any questions. And, any help is great for me!

Posts: 5,487

## Re: combine data sets fro regression

[ Edited ]

You could make your work a lot easier by changing your data format to

Data resid:

individual year city province

with one row per individual and year.

Data income:

province year income

Then you could get your summary statistics by combining the two datasets like this:

``````proc sql;
create table moves as
select
A.year as fromYear,
B.year as toYear,
A.province as fromProvince,
B.province as toProvince,
count(*) as n
from
resid as A inner join
resid as B on A.year+1 = B.year and A.province ne B.province
group by A.year, B.year, A.province, B.province;

create table moveRatios as
select
A.*,
C.income / B.income as incomeRatio
from
moves as A inner join
income as B on A.fromProvince = B.province and A.fromYear = B.year inner join
income as C on A.toProvince = C.province and A.toYear = C.year;
quit;``````

(untested)

PG
Discussion stats