|
|
| (6 intermediate revisions by the same user not shown) |
| Line 19: |
Line 19: |
| | |} | | |} |
| | <!-------------------Menu bar for metrics------------------------> | | <!-------------------Menu bar for metrics------------------------> |
| − | {| style="background-color:#ffffff; margin: 3px auto 0 auto" width="100%" | + | {| style="background-color:#ffffff; margin: 3px auto 0 auto" width="50%" |
| | |- | | |- |
| − | ! style="text-align: center; background-color:#fee5d9; width: 25%"| [[Analysis - Tools| <font face = "Trebuchet MS" color="#cb181d" size=2><b>Tools</b></font>]] | + | ! style="text-align: center; background-color:#fee5d9; width: 50%"| [[Analysis - MidTerm Analysis| <font face = "Trebuchet MS" color="#cb181d" size=2><b>MidTerm Analysis</b></font>]] |
| − | ! style="text-align: center; background-color:#fee5d9; width: 25%"| [[Analysis- Algorithms| <font face = "Trebuchet MS" color="#cb181d" size=2><b>Algorithm</b></font>]]
| + | ! style="text-align: center; background-color:#fee5d9;width: 50%"| [[Analysis - Final Analysis| <font face = "Trebuchet MS" color="#cb181d" size=2><b>Final Analysis</b></font>]] |
| − | ! style="text-align: center; background-color:#423437; width: 25%"| [[Analysis - MidTerm Analysis| <font face = "Trebuchet MS" color="#FFFFFF" size=2><b>MidTerm Analysis</b></font>]]
| |
| − | ! style="text-align: center; background-color:#fee5d9;width: 25%"| [[Analysis - Final Analysis| <font face = "Trebuchet MS" color="#cb181d" size=2><b>Final Analysis</b></font>]] | |
| | |} | | |} |
| − | <!--main content-->
| + | [[Image:Analysis.jpg|400px|center]] |
| − | <div style="background: #fee5d9; padding: 5px; font-weight: bold; line-height: 1em; text-indent: 15px; border-left: #cb181d solid 32px; font-size: 16px"><font color="#cb181d">Introduction</font></div>
| |
| − | <div>
| |
| − | For this project, our group would be focusing on the text and sentiment analysis of the Yelp reviews. We would be analysing the Yelp reviews for the Bars, BreakfastAndBrunch, CoffeeAndTea, Desserts and Bakeries that are located in the US states. We have chosen this 5 restaurant categories as they provide the Yelp users with a relatively comprehensive list of food selection. In addition, We have chosen the US states as Yelp has a lot of active users there.
| |
| − | | |
| − | After some discussion and consultation, we have came up with several questions as stated below. We felt that these questions would be useful to the restaurants located in the US. Now, lets examine these questions more closely with the Tableau Dashboard that we have built for this project.
| |
| − | </div>
| |
| − | | |
| − | <div style="background: #fee5d9; padding: 5px; font-weight: bold; line-height: 1em; text-indent: 15px; border-left: #cb181d solid 32px; font-size: 16px"><font color="#cb181d">What are some common desired attributes that customers are talking about?</font></div>
| |
| − | | |
| − | <div>
| |
| − | The common desired attributes can vary across the 5 main restaurant categories that we have chosen previously. For example, the people who visit the bars value the attribute: happy hour. This is understandable because the people go to the bar primarily to drink and they can get their beer at a relatively lower price. As such, they are able to enjoy their friends' companion while at the same time enjoy cheap beer. On the other hand, when people go to visit the restaurants to consume breakfast or brunch, they value the attribute: good food such as fried chicken etc. This is because they are hungry and need to fill their stomach.
| |
| − | | |
| − | By displaying the bigrams (adjective + nouns) in the form of word cloud across the different business categories in the US, our group is better able to understand the desired attributes that the customers value. Below are the positive sentiment word clouds for the different business categories across all 6 US states namely AZ, IL, NC, NV, PA and WI:
| |
| − | | |
| − | [[Image:Bakeries.png|400px|thumb|center|Bakeries]]
| |
| − | [[Image:Bars.png|400px|thumb|center|Bars]] | |
| − | [[Image:Breakfast.png|400px|thumb|center|BreakfastAndBrunch]]
| |
| − | [[Image:Coffee.png|400px|thumb|center|CoffeeAndTea]]
| |
| − | [[Image:Desserts.png|400px|thumb|center|Desserts]]
| |
| − | | |
| − | From the positive sentiment bigram word clouds, we can infer that it is important for
| |
| − | | |
| − | Bakeries:
| |
| − | #Bakeries to provide delicious fresh and great bread i.e. French bakery
| |
| − | #Bakeries to employ friendly stuff to provide friendly and great service
| |
| − | | |
| − | Bars:
| |
| − | #Bars to have happy hours so that the cheap beers can attract more people to drink beers there.
| |
| − | #Bars to have a great atmosphere / place by having good music, good beer, good selection and excellent service
| |
| − | | |
| − | Breakfast And Brunch:
| |
| − | #Breakfast And Brunch restaurants to employ friendly stuff to provide friendly and great service
| |
| − | #Breakfast And Brunch restaurants to provide great and amazing food i.e. coffee and French toast
| |
| − | | |
| − | Coffee And Tea:
| |
| − | #Coffee And Tea restaurants to provide great, delicious and iced coffee
| |
| − | #Coffee And Tea restaurants employ stuff to provide friendly service
| |
| − | | |
| − | Desserts:
| |
| − | #Desserts restaurants to provide great food such as hot chocolate, delicious ice and great yoghurt etc.
| |
| − | #Desserts restaurants to know the favourite flavour of its customers so they will come back again
| |
| − | #Desserts restaurants to employ friendly stuff to provide friendly service
| |
| − | | |
| − | Besides examining the positive bigram word clouds, one can also examine the positive sentiment word types such as word, adjective and noun word clouds. Word cloud is good as it tells people what are top words by its size. The larger the size, the more frequent the word appears in the Yelp Review. At the same time, the colour of the word cloud can be used to indicate the different groups of words. In our case, we decided to use the colour green to indicate positive sentiment word and the colour red to indicate negative sentiment word.
| |
| − | | |
| − | Due to word cloud limitation of comparing relative frequency between words , we decided to include bar chats to help us visualise the frequency of the word cloud. Let us take a look at how the bar charts can help to answer the desired attributes of the bakeries that are located in the NV state in the US.
| |
| − | | |
| − | NV:
| |
| − | [[Image:NV.png|600px|thumb|center|NV]]
| |
| − | | |
| − | PA:
| |
| − | [[Image:Pa.png|600px|thumb|center|PA]]
| |
| − | | |
| − | From the above picture, we can see that the top 5 positive sentiment adjectives that are associated with the chosen business categories in both NV and PA are quite similar.
| |
| − | However, there are still some differences that exist between what people value in the business categories.
| |
| − | | |
| − | Bakeries:
| |
| − | | |
| − | PA people look out more for French toast when they go to bakeries while NV people look out more for fresh food items.
| |
| − | | |
| − | Bars:
| |
| − | | |
| − | PA people look out more for friendly service while NV people look out more for nice food and drinks
| |
| − | | |
| − | Breakfast and Brunch;
| |
| − | | |
| − | PA people look out more for friendly service while NV people look out more for nice food
| |
| − | </div>
| |
| − | | |
| − | <div style="background: #fee5d9; padding: 5px; font-weight: bold; line-height: 1em; text-indent: 15px; border-left: #cb181d solid 32px; font-size: 16px"><font color="#cb181d">Why are some of the businesses thriving and others are not? What are the best practices?</font></div>
| |
| − | | |
| − | <div>
| |
| − | In order to know why the business is bad, the business owner should examine the negative sentiment word that are associated with the business. One way is look at the negative sentiment bigrams for the respective business categories across the US states.
| |
| − | | |
| − | [[Image:NegBakeries.png|400px|thumb|center|Bakeries]]
| |
| − | [[Image:NegBars.png|400px|thumb|center|Bars]]
| |
| − | [[Image:NegBreakfast.png|400px|thumb|center|BreakfastAndBrunch]]
| |
| − | [[Image:NegCoffee.png|400px|thumb|center|CoffeeAndTea]]
| |
| − | [[Image:NegDesserts.png|400px|thumb|center|Desserts]]
| |
| − | | |
| − | IL:
| |
| − | [[Image:Last.png|600px|thumb|center|IL]]
| |
| − | | |
| − | For this project, we define the good restaurants with 4.0 business rating and above while the bad restaurants with 2.5 business rating and below. From the picture, we can see that the good restaurants have a greater proportion of positive reviews compared to the bad restaurants.
| |
| − | </div>
| |
