GNOME Bugzilla – Bug 627461
time series analysis interpolation test case returns #VALUE! error
Last modified: 2010-08-20 20:18:09 UTC
Created attachment 168369 [details] time series analysis test workbook Goal. Compute asfr(time, age-group) where time is [1991,. 2021] and age-group=[15-19,20-24,25-29,30-34,35-39,40-44,45-49] Given. start and stop months of survey age-group of respondent at time of survey fertility rate based on number of births in the three year period befor interview Imputations. Start and stop dates are the 15th of the month of the dates of field work. Interview conducted at the midpoint of the start and stop dates Fertility observation ascribed to 18 months before the midpoint Data observations. Survey name vlookup of date from the survey-date worksheet (abscissae) births per thousand by survey and age-group (ordinates) Output.asfr by calendar year and age-group Formula. =interpolation('asfr-dhs'!$B$2:$B$6,'asfr-dhs'!$C$2:$C$6,'asfr-ts'!$A3,0) where 'asfr-dhs'!$B$2:$B$6 are the imputed dates of the observations 'asfr-dhs'!$C$2:$C$6 are the observations 'asfr-ts'!$A3 is the target date 0 is the method
Created attachment 168370 [details] updated workbook
Created attachment 168374 [details] updated updated workbook Close this bug. I made two mistakes: 1) the given abscissas must be in increasing time order 2) the returned values must be of type number and not date
Well, we still need to fix the documentation to point out that the abscissae must be in increasing order.
This problem has been fixed in the development version. The fix will be available in the next major software release. Thank you for your bug report.