× Limited Time Offer ! FLAT 20-40% off - Grab Deal Before Its Gone. Order Now
Connect With Us
Order Now

DATA4200 Data Acquisition and Management Report 1 Sample

Assessment Description

Business Problem: Airbnb is a U.S. company which provides an online marketplace for short- term and/or holiday accommodation. Airbnb collect large volumes of data to gain insight into their clients and associated customers, such as review scores, host acceptance rate, ‘superhosts’, popular accommodation types and density of listings in particular location.

Data sets: We have obtained data on Airbnb listings in Melbourne with a variety of variables. Sampled datasets, the original data and data dictionary will be available from Week 4. See sections below.

Assessment Instructions

Analysis and Report

Use Microsoft Excel or Power BI or Tableau.

Recall the sampling methods below that you have learnt about in lectures.

• Non-probability sampling

• Probability sampling

A data dictionary file and the following datasets (as .csv files) that contain sample data generated using quota, systematic, simple random, and stratified sampling will be available from week 4, see section c. below. You will also have to access the original population dataset cleansed listings dec 18.csv from the source, see section a. and section e. below.

Create a report and include your response to the following questions:

a. Access the data file cleansed listings dec 18.csv, by going to the link provided on MyKBS under the Assessment 1 tab. You will initially be downloading a zip folder from the Melbourne Airbnb Open Data project on Kaggle. Extract all the files within the folder and then choose the file cleansed listings dec 18.csv. Browse over the columns and comment on which variables appear to be the most useful in terms of insights into current listings. Document that in your report.

b. List an advantage, possible disadvantage and limitations of each of the sampling methods.

c. Access the sampled data sets on MyKBS. Choose a number of different variables, as in part (a), then for each of the sampled datasets create summary statistics for each of those variables. That is, make sure that the selected variables are the same for each of the four datasets and document them in your report.

d. Interpret and compare the results of the summary stats across all four sample datasets. What conclusions can you draw from the comparison. Document your findings in your report.

e. Repeat the above for the original dataset cleansed listings dec 18.csv. Explain with statistical examples which sampling method summary stats (across all chosen variables) were nearest in value to the original dataset summary stats.

Explain the variations in your report and include the supporting data. Explain possible ethical issues that could occur from the use of sampled data.

Briefly evaluate the software that you have used to produce the summaries.

Solutions

These 34 variables will be used to form an organized and meaningful insight. The rest of variables will remain out of the scope for this analysis. The above variables will be used to identify Airbnb location, details of the host, what is the type of the property, number of rooms, bathrooms and bedrooms, so that the customer can have all details about the property before booking.

b. List an Advantage, Disadvantages and Limitations of the Sampling Methods

Non-Probability Sampling

Quota – Quota sampling is based on participant’s quotas or characteristics using some demographic factors like age or gender. The recruitment is done until each quota is met. The main advantage of quota sampling is saving of time as well as resources. This sampling is more straightforward and minimum resources are required to conduct this sampling.

Disadvantage – As the selection of participants is based on particular quota which is ease to access and low in cost, that results in sampling bias for the assignment helpline.

Limitations – As the sampling is not based on random selection, this may be a reason for research bias.

Convenience – This is the easiest and quick method for data sampling that are cost effective and easy to reach.

Disadvantage – This also cause sampling bias as this is based on unfair advantage to certain community.

Limitation – Accuracy of data is at stake as only easily contacted participants are reached to collect data.

Probability Sampling

Stratified – This sampling ensures that each stratum is represented accurately and also helps in minimizing sampling error as well as variability.

Disadvantages – This sampling can be difficult.

Limitations – Its difficult for researchers to confidently classify every participant into a subgroup.

Systematic – This technique of sampling can eliminate the clustered selection and also there are less chances of contamination of data.

Disadvantage – This sampling can provide inaccurate results because of unavailability of some data.

Limitations – This sampling may cause over or under representation of some particular pattern. Also, there is a greater risk of data manipulation.

c. Creation of Summary Statistics

We have selected two variables from the above list to develop the summary statistics. These variables are price and review score rating. Various statistics values are calculated using Tableau and MS Excel. Different types of functions are used to calculate all related statistic values. The average value of room price is 167.98 and the maximum price value is 162.67 whereas the minimum value is 14 for price variable. The similar stat values are calculated for review score rating.

Stratified Sampling

Quota Sampling

Random Sampling

Systematic Sampling

d. Comparison of all four sample data sets

Two variables have been considered to draw the comparison between given data sets. These four data sets include systematic, stratified, random and quota sampling techniques.

As we know that stratified sampling is more accurate as compared to other three, so average of price is strategically accurate in this sampling. On the other hand, standard error is mainly used to measure the efficiency and accuracy of the sample, but it has large value for stratified sampling. This implies that the stratified sampling dataset consist of discrepancy. There is no change in the values of median and mode. Standard deviation shows the measure of dispersion of data. From the comparison, it is clear that stratified data has maximum value, that indicates how wider the data is spread out. Hence, it signifies higher variability. Sample variation can be defined as a measure of dispersion of observations around the mean. Stratified sampling shows higher value of the variable that indicates the collected data can have higher variability. On the other hand, quota shows lower value of std deviation that indicates similar data collection.

Kurtosis defines the degree of peakedness whereas Skewness is degree of asymmetry of the distribution. Random data has higher value of kurtosis that shows greater probability of deviation from the mean. Stratified data is on second number. For all four data sets, distribution is positively skewed.

As the above curve is lower in peak, therefore, it is known as Platy_kurti. It shows that the chosen variable will be more stable as compared to other variables. For skewness, the curve is positively skewed.

The second variable chosen for summary statistics is review score rating. Quota sampling shows maximum value for mean whereas standard error has largest value for stratified data sampling. Similarly, maximum value is derived for standard deviation statistics. Skewness is negative for all data sets which show data for this field is highly skewed.

e. Using the Original Dataset

The same step is repeated for calculating summary statistics for cleansed listings dec 18.csv. All above data sets are merged into a single data set and then cleaned by removing null values and adding some values to the price field that consist of 0 values. All fields like Mean, standard error, median, mode, standard deviation, variance, kurtosis skewness and so on are calculated by using the same functions and by similar methods. The file consists of approximately 22000 data elements so its processing remains slow as compared to other data sheets. The average of price calculated is 148.07 which is in between of previous values. Standard error values is 210.84 which is still large, that indicates about discrepancy of the data set. Median and mode values are equivalent to the four data sets. Not much change has been notified in these values. Again standard deviation value comes to 210 showing the wider spread out of the data. This indicates that data is not limited to quota, systematic and random sampling. More data sampling means many more visualization can be possible and it leads to the perfect data analysis that may affect positive decision making by the company. Skewness is positive and its value is 26 which is a large value that indicates an excess of low values for the price field. High value of kurtosis shows distribution is more peaked. The summary statistics table for the price variable is shown below.

The summary stat calculations for review score rating variable is listed as following. According to the table, the average value of the merged field is 94.18 which lies in between the value calculated for all data sets. The discrepancy for this column is less because of small value of standard error. The calculated sample variance value for the field is 73.40 which is not much high and indicates less variability of the collected dataset. Again, kurtosis value is lower and the result of skewness calculated field is negative.

From the above discussion, by calculating summary stats for all four data set and the cleansed data set, it is clear that the value of that related field remains equivalent as both variables consists of similar values in all sheets. The only main difference lies in speed of processing of both types of data set.

Fill the form to continue reading

Download Samples PDF

Related Sample

Assignment Services