geography f764 - skills CBD investigation

Alternate hypothesis 

'How and why do rateable values, pedestrian counts nad land-use vary with distance from the PLVI within Ipswich's CBD' 

• 'Pedestrain flows decrease with distance from PLVI' 
• 'Rateable values decrease with distance from PLVI'
• 'Land-use changes with distance from PLVI with an increase in % of convenience stores' 

BROAD TOPIC 

• Teachers input on practicality / feasibility 
• CBD - urban change + structue of urban areas 

LOCATION (Ipswich)

• easy access / no cost / few hazards / local knowledge 
• lots of variables to test 

EXACT SITE AREA (Goad)

• primary + secondary shopping / including PLVI 
• suitable scale / testable area 

HYPOTHESIS 

• geographical theories - BID RENT / THRESHOLD POPULATION / DISTANCE DECAY 
• SMART 

GOAD 

• covered sensible range size for data collection + suitable detail scale 
• covered primary + secondary shopping areas 
• doesnt show zone of advancement south 

PRIMARY - PEDESTRAIN FLOWS 

• 30 selected sites - reliability + accuracy (representative) 
• point data 
• both directions, just walking, for 2 mins 

PRIMARY - LANDUSE 

• whole goad 

SECONDARY - RV 

• 30 to keep data comparable + accurate 
• Valuation Office Agency Site 

SECONDARY:

• GIS / teacher / text books / stat tests / RV / theories / hypothesis 

• test resources / equipment + adjust 

• pre test data collection methods (sampling frames)

• see if it can be done in given time + location 

• identify hazards / problems 

PEDESTRIAN FLOWS 

• point 
• 30 sites 
• systematic / pragmatic 

LAND USE 

• no sampling frame 

RV 

• point 
• 30 sites 
• systematic - comparable 

RELIABILTY 

• only 30 sites 
• sytematic - bias 
• counts at different times of the day 
• no repeat 
• only 2 minutes 

ACCURACY 

• pedestrian flow count - human error 
• land use - not every building in each class 
• weather 

• postcode of sites entered onto Valuation Office agency website 
• of 30 buildings (comparable to Pedestrian flows)
• systematic - even spread 
• Goad maps to find distance frm PLVI 

LIMITATIONS 

• secondary data (cannot be certain of validity)
• standardised so size will not affect 
• not all adresses had RV 
• ruler used on Goad map - HUMAN ERROR 

• primary - reliability 
• 30 sites as respresents statistical population, acheivable, statistically tested
• systematic - evenly covered at every 8th building 
• ensure DISTANCE tested - 4 groups 
• 2 minutes of recording 

LIMITATIONS 

• systematic - bias 
• counting spread out over 1.5 hours 
• perpendicular streets to PLVI - unsure if going to or away from PLVI 

• primary data - classification index (incl. comparison + convenience)
• pilot study - whole statistical pop. of Goad could be covered + hazards identified
• groups 1-9 including primary and secondary shopping areas 
• including PLVI + 'GOLDEN MILE' 
• including Westagate st, Tavern st, Carr st 
• split Goad into 4 sections for each group 

LIMITATIONS 

• tough to classify each building 
• larger stores could be both department and comparison 
• Goad map created in year 2000 - now new buildings + shops (e.g. end of princess street estate agents + pub) 
• Goad map created in year 2000- buildings knocked down 

CHLOROPLETH MAP 

• visually display RV of 30 buildings 
• categories chosen evenly across range of RV 
• chose 6 categories 
• 'no data' category created 
• simple + made good use of available resources 
• light to dark shade depending on value 
• each areal unit concealed + abrupt changes 

SCATTER GRAPH 

• present in relation to distance 
• distance was independent variable and RV were dependent variables 
• simple and showed basic correlation 
• see if stat. test should be carried out 
• only suggests trend with no figure 

ISOPLETH MAP 

• easy to construct and pick out surface trends 
• data added to isopleth, so no abrupt changes 
• colouring scheme of light and dark for low and high values 
• 3 classes to keep simple and effective 

SCATTER GRAPH 

• to test distance with pedestrain flows 
• clearly show points and how close to line of best fit 
• examines relationship between 2 variables 
• estimate correlation to see if further stat. tests need to be carried out 
• anomalous points skew data 
• 30 is max as becomes confusing 
• only relationship between 2 variables 

DIVIDED BAR CHART 

• used to show the 9 variables (classes) on each bar - so hypothesis can be tested 
• created new index - removed grey areas in defining buildings - reducing HUMAN ERROR
• each bar was 0-100m of 300m section (3 bars)
• simple and allowed enough comparison 
• could have been split into more for more accuracy (would have been tougher)
• coulour scheme allowed visual representation 
• data recovered for stat tests 
• could have used 3 pie charts instead (visual comparison and % for stat analysis) 

CHLOROPLETH MAP 

• mulitple sites along 'GOLDEN MILE' are highest rate even with increasing distance 
• outside this rates drop off suddenly 
• supports hypothesis once passed Westgate St + Tavern St 

SCATTERGRAPH 

• correlation overall supports the hypothesis - trend line 
• Westgate (highest) UPPER ORWELL /CARR (lowest 

anomalies 

• Tower Ramparts - bus stops - main entrance
• St Matthews St - not pedestrianised - rapid decline 

STAT TEST 

• -0.5 - 99% confidence level therefore can be very sure that sample is representative of total statistical population and isnt due to chance 

ISOPLETH MAP 

• 'GOLDEN MILE' has more of an inflence on pedestrain flows rather than single PLVI point
• rapidly declines around golden mile 
• different comparison shops and access points (car parks) affect p.flows 

SCATTER GRAPH 

• negative correlation supports hypothesis 
• -0.46 passing significance test at 95% confidence level 
• time and p. flows has 0.329 passing significance test at 90% level (had some affect)
• RV + p.flows has 0.897 passing at 99% confidence (strong relationship) 

ANOMALIES

• low - St. Matthews st - west of PLVI - not pedestrianised
• high - Tavern St. + PLVI - 'GOLDEN MILE' 
• tower ramparts - vacant stores + redevelopment 

DIVIDED BAR CHART 

• overall with distance from PLVI % comparison stores did decrease and % convenience increased 
• 0-200m comparison increased a bit, then rapidly decreased 200-300m 
• 0-200m convenience decreased to 0, then rapidly increased 200-300m

DIVERSITY INDEX 

• a stat. test that allowed us to see diversity increase with comparison stores, showing lower % of that store with increasing distance 

VACANCY RATE - Upper Orwell - 0.529 = zone of discard + very low retail land value 

ANOMALIES

• unexpected high value between 100-200m for comparison stores due to location along 'GOLDEN MILE' is close to Tower Ramparts + walkway to Buttermarket 
• size of comparison store affected count - 1 big one will look less than lots of small 

• decline with distance 
• support BID RENT + DISTANCE DECAY - 99% confidence level 

other factors 

• west of PLVI - not pedestrianised 
• values affected by who is resposible - owner occupied / reapirs / insurance 
• not standardised values used - building size had a MASSIVE effect 

• accept alternative hypothesis 
• negatove correlation with 95% significance 

OTHER FACTORS

• time positive correlation 90% confidence (only measured over 1.5 hr period) 
• RV poistive correlation 99% confidence - larger buildings higher RV (not standdardised) 
• routed distance would have produced stronger relationship over straight line 

• accept alternative hypothesis 
• general trend that compaison stores do decrease with distance from PLVI 
• GOLDEN MILE has most comparison - higher THRESHOLD POPULATION 
• DISTANCE DECAY - as vacant stores increase 

OTHER FACTORS 

• TOWER RAMPARTS - caused higher surrounding of comparison - access point to other shops