Data Collection at Moshi Office - Weir + Swai, 15/9/94

1. Approx 68 files collected from hard discs and from various floppies - these are raw data files - .dbf database Files but NOT in dbase3+/dbase4/clipper format -- will have to write extraction program to process these into useable files; my comments and feelings on this data confirmed independently by Kilimanjaro computers of Moshi, who looked at the system for Martin Doran at one time; the system is ENABLE - a database system written I believe by the makers of Ability Integrated package

2. These files are valid for Arusha Sept 91 thru Aug 94 with gaps at 94/04 (no data collected?), 92/04 - 92/11, 93/04 -- 93/05, and 93/07.

Also for Moshi 91/08 thru 84/08 with gaps at 93/04, 93/06, 93/07, and 94/04.

All most gaps filled by photocopying 240/ 250 original data sheets

3. The sheets also had on the header the number of beasts yarded and the number sold for each week; this info was input by me manually for 300 data points ( 3 years x 52 weeks/year x 2 locations)

4. Problem - these files each contain data for location and month, but within each file the week is NOT specified. Thus we may need to recapture that info from the original paper info; in fact, we can assume that data is entered in week sequence, with sample size proportional to yardage.......

Data Collection at Moshi Office - Weir + Swai, 17/9/94

1. Crosschecked on missing data - April 94 Moshi and Arusha still missing - people on leave etc.. - was not collected...

Arusha 30/4/93, 8/10/92 and 18/9/92 all missing - no great problem

2. Read Technical End of Mission Report, Martin Doran , Dec 1993

- Korogwe data quality poor

- Doran stated that Swai was good (data processor),

Ali good (data collector Temi) but Kilimba poor (data of erratic quality for

Weruweru)

- Doran referred to weight and price data analysis from Trekking report

- 1990 report FAO referred to 140 single-beast weighing units being donated to Tanzania; Swai had no knowledge of their whereabouts

3 - Read URT - Minag "Northern Trekking Routes and Terminal Markets" 1991 FAO/MDB - WP1 /1991 February 1992

pps 24 & 25 give graphs of price vs yardage

4. Read accompanying/later report: ""Volume 2 -- Livestock Production and Trekking Route Surveys - policy Implications and Recommendations" FAO/MDB

Nov 1993

pps 37-47 and pps 85-88 all highly relevant to MIS project.

- volumes at Themi and WW fluctuate with season - single annual cycle peaking May-Aug, troughing Oct-Jan

- good correlation (inverse of course) yardage vs. price /kg

- price /kg fairly independent of grade, but report not specific as to whether live weight or carcase weight (I guess carcase weight price more constant across grades)

- high quality cattle mainly sell at Themi, unsold go to WW

- unofficial exports to Kenya discussed, sometimes Kenya butchers will buy OUTSIDE Themi to avoid permits/detection

- price/kg in US dollars is declining slightly with time (actually report examined price in constant Tsh corrected by official retail price index; which may be skewed?)

- Themi is underreporting yardage and sales; irregularities are suspected as one reason for this....

-- detail on various fees is given in some detail on footnote page 37

5. This volume 2 report refers on page 42 to the 1991 report Appendix A Section 1 number 9 for some data - cannot find this Appendix or data, which should be quite relevant -- must find this through MDB.

6. Checked data structure of the .dbf files copied -

Header approx 1014 bytes,

1- 10 origin

11-15 steer/bull/cow

16-17 age in years

18-18 grade

19-21 weight in kg

22-27 price in Tsh

28-28 blank

next record starts on byte 29

7. Contact data -

Mr Swai, Livestock Field Officer, Mifugo REO, Moshi, tel/fax Moshi 51229, telex 43005 KAIDC TZ

8. Helped make space on hard discs of both PC's; chkdsk'd both PC's (lost clusters found on 8086), ran NDD on both, Speedisk on 286, tried to install mouse - 8086 has no com port, mouse eventually found to be inoperational; generally set up both PC's using autoexec.bat and config.sys to be ready for he new MIS system, so can be installed by Swai from disc without site visit by Weir

Strongly recommend that both machines be upgraded - at least replace 20 meg hard discs with 100 meg+; also consider main board upgrade to 386DX 40 Mhz or similar with 1 Meg Ram (US$ 300-00 approx duty paid price per machine); hard disc upgrade I guess at also in region of US$ 300-00 per machine. As afterthought of course compare costs with replacements.....

9. Copy of Volume 2 report as above borrowed - to be returned by TLMP at earliest