Please help on question 3 and question 4, link of data source as following, thanks!

Data Source:

https://docs.google.com/spreadsheets/d/1G1g37WUjRLn9a-LYHpIIefFUozuz6ncj_DF9jEtN9Jo/edit#gid=112983759

https://docs.google.com/spreadsheets/d/1G1g37WUjRLn9a-LYHpIIefFUozuz6ncj_DF9jEtN9Jo/edit?usp=sharing

Meeting Purpose: Specifying and Allocating Data Analytics Tasks Discussion Model Quantity Ordered. items: Model the likelihood of recommending B-Hive to others. Forecast demand for the upcoming four quarters. Produce a technical report. Detailed Who: What: Action Items 1) Build a multiple regression model to estimate the order quantity. 2) In a separate project, Peter found that the perception of product quality is a significant predictor of the quantity ordered. In line with his findings, prior research shows that the strength of this relationship may vary according to brand image. That is, customers tend to associate the brand image with product quality. Therefore, Peter believes that the relationship between quality and quantity ordered should be stronger for those with more favourable brand perceptions. Model the interaction between image, quality and quantity to test Peter's assumption and comment on whether there is sufficient evidence to conclude that the interaction term is statistically significant in the model. 3) Build a logistic regression model to predict the likelihood of recommending B-Hive products to others. 4) Peter has completed the first step for this task. He has narrowed down the key predictors of the likelihood of recommending B-Hive products to "Distribution Channel, Quality and Brand Image". a) Continue Peter's work and develop a logistic regression model to predict the "likelihood of recommending B-Hive products". b) Create a plot that shows how the changes in perceptions of quality (scores from 1 to 10), a positive brand image (a score of 10) and the distribution channel (i.e., those purchasing directly or through a distribution network) affect the predicted probabilities of recommending B-Hive products.The Case Study The B-Hive is a regional victorian honey producer with a 20-year history. Although its operations are limited to the Gippsland area, the company has been financially successful. B-Hive honey is sold directly to customers {Public, lGrocery Chains, Shops and Restaurants} or indirectly through an external distribution network. Despite its succasful operations and solid financial turnovers in the last two years, BHive forecasts a shift in consumer demand. Now more than ever, BHive management feels the need to ensure a strong relationship with its diverse customer base. The management team wants to understand the characteristics of BHive's customers and their repurchase intention. In addition, they would like a formal procedure to forecast demand for their honey. The formal forecasting would help BHive with demand planning and production scheduling. Data The Data and Insight team conducted an online survey asking customers to rate the company on nine attributE using a 1 10 scale. The customers also indicated whether they would recommend B- Hive products to others. The survey data was supplemented by information from BHive's databases like the loyalty duration, customer type, region, distribution channel and other information. A complete listing of variables, their definitions, and an explanation of their coding is provided in the 142122022.fo file. Peter [your team leader} has allocated relevant research tasks and explained his expectations from your analysis in the meeting. The minutes of this meeting are available on the next page. Your task is to review and complete the allocated activities as per the document. Iask a Model building It is IMPORTANT to follow an appropriate modelbuilding process. Include all steps of the model building activities {especially all relevant pre and post-model diagnostics} in your analysis. Watch out for indications of oollinearity [look at the correlation matrix ,I' large standard errors} and numerical instability {complete separation! nearcomplete separation}. The main aim is to identify the independent yariab les that impact group membership of the dependent variable. Include many Excel worksheets [tabs] as you require to demonstrate different iterations of your regression model {i.e., 3.1, 3.1.a]. Please note all reasonablefrealistic assumptions about the parameters next to the analysis. Task 4.a & 4.b - Visualising and interpreting predicted probabilities To accomplish this task, develop a new logistic regression model using ONLY the factors discussed in the team meeting (Item 4). Then use the model to develop a predicted probability plot using the parameters mentioned (Item 4). The technical report must include the predicted probability visualisation and practical recommendations. These recommendations should broadly answer the following question: "How changes in perceptions of quality (scores from 1 to 10), a positive brand image (a score 10) and the distribution channel (i.e., those purchasing directly and through a distribution network) affect the predicted probabilities of recommending B-Hive products."