Big data analytics, data mining, machine learning, and finding product improvements for exploding points.

In order to avoid unnecessary trouble. Some of the data is not very clear, the key to see the point of thinking.

Through statistical analysis of big data, I found that a linear formula can be used to perfectly fit a user conversion link. Based on this formula, we make predictions about the data that haven't occurred recently. That's the beauty of big data.

According to the above findings, I think at least we this product, the Internet funnel model, in every two links between the transformation model should be linear model. In turn, continue to use the data to explore another transformation link. The results found very sad reminders.

It is completely impossible to fit with a linear formula. Even disorganized. Wait a moment. Is there something wrong with it? So I'll continue to explore the relationship between referrals and conversion rates.

Mom, the more referrals you find, the lower the conversion ratio. It's not scientific. At least it should be a fixed proportion to say the past. This inverse model is too much, and it's going to waste a lot of traffic.

Continue to explore page click Pv/uv Analysis (Skip) ...

Finally find a friend circle to share and collect the hourly data graph

The results found that the friend circle limit flow, basically share the number of times a 15,000 is dry down. After July 14, it is completely limited to the peak of the current level.

Through the above analysis, we find that the bottleneck of our system is the limit flow of the circle of friends. Solution business negotiation, or multi-domain. Is there any other way, if the great God knows. please tell me. Thank you first.

Big data analytics, data mining, machine learning, and finding product improvements for exploding points.