"Big Data Application Scenario" Next Door Pharaoh (Series 3)

The last time Lao Wang made a big play with the help of Xiaobian was not successful, but he was smothered in the cradle by employees with the threat of pa gong. For Xiao Wang's strong growth, Lao Wang has to honestly predict the sales volume for next year.

Lao Wang's inner drama: Now that a certain dragon spicy tiao is going international and is deeply loved by people from all over the world, what can my spicy tiao factory do to become the second certain dragon? Lao Wang couldn't help but fell into contemplation. What has something to do with sales? Taste? shape? What kind of spicy sticks do you recommend? The taste is out of the question, so we can only start with the shape and the main model. Thinking of this, Lao Wang excitedly ran over and smashed the door of Xiaobian's company, looked at Xiaobian with expectant eyes, and started a new round of bitterness. begging.

As before, Lao Wang excitedly stated his needs. After the editor helped him sort out the general process, he began to make sales forecasts (if there are good neighbors like us, please give me a dozen, they are all free. ...heart is dripping blood)

Step 1: Collect data through Forespider

First of all, Xiaobian first searches for spicy bars on Taobao, Tmall, Yihaodian and other major e-commerce platforms, and uses Forespider to check the sales and all information about various spicy bars. Collect comments.

Step 2: Mining the comment information through the data mining system

After collecting all the data, it is found that the comments are numerous and complex, and there is no way to intuitively see the buyer's preferences. Function, use keyword search, set keywords including taste, packaging, raw materials, taste and so on. From this, separate reviews containing flavors, packaging, ingredients, flavors.

Later, the sentiment polarity analysis of the classifier, also known as text orientation analysis, is used to judge whether the comments in the comments are delicious, unpalatable or ordinary. Through this judgment, Lao Wang got a batch of delicious comments, as well as the corresponding products and product sales.

Step 3: Use ForeAna to conduct sales statistics and analysis

After the data was collected, Lao Wang was stunned when he saw the sales of various types of spicy sticks. . . The editor told Pharaoh, don’t be afraid, forget that we still have the ForeAna analysis system (when it comes to this editor showing a profiteer’s smile), and then associate the classifier with the ForeAna data analysis engine to automatically obtain information about sales, shape etc. visualization charts.

Step 4: Build a Mathematical Model

But what Lao Wang has to do this time is to predict next year's sales, so the above information and steps are not enough. Therefore, Xiaobian began to model based on the collected data results, using the sales volume obtained as the dependent variable, and the taste, packaging, and raw materials as independent variables to establish a mathematical prediction model.

Step 5: Obtain the results of the analysis model

After the model prediction and analysis, the editor successfully predicted the sales of various spicy strips for next year for Lao Wang: 4 million in strips, 3.5 million in blocks, and like 3 million. Lao Wang finally realized the drawbacks of his own production of single spicy strips, and secretly thought that the sales of strips were the best, followed by lumps, followed by flakes. Excited to shout that next year will not only continue to produce block spicy strips, but also increase the production of strips and flakes!

Out of gratitude, Lao Wang went back to the company and brought a box of spicy sticks in person, so that the editor could eat enough and sprinkle flowers~