You are on page 1of 25

Analytical Method

I.1. Introduction
DATA Usage: Conclusion, Decision Data requirements : valid, representative DATA condition : - several institutions collect similar data - data are available in some institutions - different data form - reliability ???

Traffic Data
MONITORING The collection of prevailing information at any time and as they change over time. FORECASTING The use of current and previous data as one of the inputs for prediction of future condition CALIBRATION The use of traffic data to estimate the values for one or more parameters in a theoretical or simulation model VALIDATION verification of a theoretical or simulation model against information independent of that used to calibrate the model.

Preliminary planning

Survey form design Pilot survey Survey administration coding

Sample design

editing Analysis Report result

System Approach to Traffic Analysis Archive survey and data

PRELIMINARY PLANNING
Study objectives Survey objectives Review existing information Formation of hypotheses Definition of parameters Resource determination Survey content Selection of survey methods

SAMPLE AND EXPERIMENTAL DESIGN


Definition of target population, sampling unit Selection of sampling frame Choice of sampling method Choice of experimental design Consideration of sampling error and bias Determination of sample size

SURVEY FORM DESIGN


Factors to be considered: Ergonomically designed Weather resistant Convenient size and good format Sufficient space Survey forms should be tested Provided with a clear and convenient set of instructions

PILOT STUDY
Provide assistance: Adequacy of sampling method Variability of parameters in the survey population Suitability of survey & analysis method Adequacy of survey forms Efficiency of training methods Suitability of coding and editing Cost and duration of survey and analysis Efficiency of organization

SURVEY ADMINISTRATION
Survey manager/designer survey administration to decrease field problems Check list: - Supervisor should arrive early at survey site - Adequate wet-weather and other disturbances protection - Surveyor replacement - Enough survey forms - Adequate rest and meal breaks

CODING & EDITING


CODING : process of converting data from field records into computer-readable format EDITING : process of scanning data to detect recording errors and logical errors

DATA ANALYSIS
Procedures : Data investigation data scanning, examination of data distribution, tables, diagrams Hypothesis testing to compare the distribution of two or more data sets: chi-squared, nonparametric, ANOVA, distribution fitting Relationship determination multiple linear regression, stepwise regression, time series analysis, etc. Data reduction large set of variables more manageable set

PRESENTATION OF RESULTS AND ARCHIVING


The importance of archiving: - it is needed by other researchers - as secondary data - good report format is necessary. Survey report principles: - Fully explain of the purposes and scope of the survey - Type of readers should be known - Language and statistical technicalities are conform to the readers - Qualitative description and graphs are required - Report types: preliminary report, summary report, technical report.

SAMPLE
Reasons of sampling: The inability to record information for the entire population restrictions: time, fund, personnel, etc. The scarcity of data, the rareness of certain events impossible to obtain a total population

FACTORS AFFECTED IN SELECTING DATA COLLECTION METHODS


Cost and ease of use Time required Number of data items able to be collected simultaneously. Accuracy of results The medium of the data collected. (e.g. computer readable, hard copy)

I.2. DATA TYPE


1.

Level of measurement - nominal : lowest level, identification/ classification - ordinal : ranks data in some order - interval: the distance between each category is defined in terms of actual units. 0 (zero) has a specific meaning - ratio : have all the properties of the above scales, as well as having definite zero. 0 (zero) does not have any meaning

Measurement Scale Nominal Scale Non-Interval Scales Ordinal Scale

Characteristics of the Scale A scale that measures in terms of names or designations of discrete units or categories A scale that measures in terms of such values as more or less, larger or smaller, but without specifying the size of the intervals A scale that measures in terms of equal intervals or degrees of difference but whose zero point, or point of beginning, is arbitrarily established A scale that measures in terms of equal intervals and an absolute zero point of origin

Statistical Possibilities of the Scale Can be used for determining the mode, the percentage values, or the chi square Can be used for determining the mode, percentage, chi-square, median, percentile rank, or rank correlations Can be used for determining the mode, the mean, the standard deviation, the t test, the F test, and the product moment correlation Can be used for determining the geometric mean, the harmonic mean, the percentage variation, and all other statisctical determinations

Interval Scale Interval Scales

Ratio Scale

2. Continous discrete data


-continous data -> variables for which possible values have no breaks, example: speeds and headways -discrete data - > variables whose values have breaks or jumps contoh : number of accidents, number of vehicles.

3. Quantitative qualitative data


-Quantitative data -> can be expressed numerically, e.g. speed, the number of vehicle, income. -Qualitative data - > are expressed by some non-numerical property, e.g. vehicle type, perception to traffic noise

4. Stationary-time dependent data


- Stationary data offer the possibility of repeating observation as experiments, e.g. stability data, asphalt mixture density, etc. - Time dependent data come from processes that are

undergoing continual change, do not permit repetition of observation, e.g. traffic volume data (consisting four components: TREND, SEASONAL VARIATION, CYCLIC VARIATION, RANDOM COMPONENT)
TREND (Tr): long term change in the average quantities, e.g. traffic growth over time SEASONAL variation (Se): different levels of flow at different times of a year, e.g. level of recreational traffic on a rural highway CYCLIC variation (Cy): result from cycles in activity (such as time of day or day of week influences on working or shopping behaviour) RANDOM component (Ra): short-term variations in behaviour (e.g. influenced by weather), or special events

I.3. FAULTY DATA


A typical fault in some data: a. outliers: data points that do not appear to be consistent with the general body of the data - necessary to return to the original data set - Is there any errors in recording ? Yes data point can be neglected No carry out the analysis with, and without outliers - If they have great influence on the result this should be noted. b. missing data , have two aspects of concern : specification and data analysis

ATTENTION IN DATA EDITING


A. Permissible Range Errors Errors in recording and typing may lead to obvious errors, where the code value for a data item falls outside the range of codes used require comparison of data set with the original recording sheets. Example : number 1 mis-read as 7, while the range of codes is 0 or 1

Logical Errors
the logical relationship between two data items for the same observed unit is not correct. Example: Truck two axles is noted as 3 axles errors identification cross tab. Checking: - as a result of recording -> the error is easily rectified - as a result of observation -> the error is examined and compared to the other item recorded to find out the inconsistent item -> correction

Missing data
There are two types : 1. completely missed - e.g. observers attention is distracted by some other event or a vehicle is obscured - correction on the basis of weighting the observed data in terms of some known characteristics of the population 2. partly missed - ignore the missing values and report - report the proportion of missing values for each variable, so that the results are based on the total number of observation made -estimate a probable value of the missing item based on the information available from the other data items for the observation

You might also like