Using data to speak: Data behind Beijing house price data

From 2014 to the widespread decline in the property market, to 2015 price warmer, in the end what happened to change? This article tries to show you the data behind the data by using big data and rich charts.

Data Sources

These data are the author in October 2014 and October 2015 two times, in the chain home online crawl of the second-hand housing data, 2014 about 64,000, 2015 total about 7W. The data source may be biased, so the conclusion is for informational purposes only.

Soaring Rates

From 1992 to 2015, how did the price of Beijing go insane? You can view the chart below.

As you can see, 1992 to 2002 presents a very stable state. Since 2008, house prices in Beijing have been as common as rockets.

Interestingly, if the chart is plotted according to the construction time, it will be found in 2000 and around 2004, reaching peak. In the 6W set of secondary housing, a total of 7697 units were built in 2000, accounting for 11.21%.

By the year 2014, the prices of housing in various districts of Beijing such as:

The average price in the Xicheng and Dongcheng District is around 50,005, and the reason for not reaching the 90,000 of other data on the Internet is because we are analyzing the secondary housing. At present, the number of new buildings in the two ring is very small, with little discussion value.

We will be very intuitive when we draw the house price on a map in the form of hot plots:

The darker the color, the higher the price. In addition to the west side, east side of these Old Town, Zhongguancun (including a large number of school district room) and the International Trade (Beijing CBD) are high prices.

If we change the zoom level and further reduce the map extent, you can see the most expensive houses, concentrated in Xidan, South Percussion Lane, Guomao, and Beixinqiao area.

How expensive is the price of these most expensive neighborhoods? The price of the top ten 10 communities is listed below:

Where is the location of Mandarin Hutong? I specifically went to search for a bit. This super bull, price in 330,000/Ping Mandarin Alley near downtown Mouth Street of Xidan shopping district.

More exaggerated, the two houses are bungalows, the area is 12 and 15 Ping, one of the set was built in 1949. Intermediary gives the propaganda slogan is, the most cattle experiment Secondary School District room , Rob Rob Rob! Such a small area, it is estimated that the courtyard of the room to transform it. Now the official online has been off the shelf.

What kind of house is the most?

We look first, what type of building the most, below gives the proportion of the overall height of the building. Can see, the secondary housing, six floor is the most. The state stipulates that the seven-storey elevator will be installed. So before 2000, most of the residential buildings were 6 floors.

Then look at the proportion of the total number of houses in different areas. We took an area of 40-140 square meters of the house, carried out statistical analysis, the conclusion such as:

It can be seen that the 60-level one-room or two-room and one-bedroom rooms are most common. 90 square meters of three and two living also more.

We will draw a thermal map of the second-hand stock:

Can be seen, the secondary housing mainly concentrated in Tian Tong Yuan, North Court, Wangjing, ten miles fort and Tongzhou. This is basically in line with Beijing's 2004 "Beijing City Master Plan 2004-2020" content:

It is interesting that the green area is quite empty, for example, the author is currently located in the three-dimensional bridge area, and Jiuxianqiao separated a good big wasteland, night run in the deserted.

changes in house prices from 2014 to 2015

The following is what the reader is most concerned about, 2014 to 2015, the price of Beijing, how has the change? As we all know, in 2014, house prices fell, the overall decline, even the business has played a price reduction of 6000 yuan/flat advertising to promote the house. 2015, the reduction of quasi-interest rate cut policy, the ratio of Provident Fund loans increased, Beijing housing prices warmer, we know that certainly the price. But how much has it risen?

We plotted the ratio of the price change in the following chart according to the percentage increase or decrease in prices for 2014 and 2015. We can see that the change of house price basically shows normal distribution trend. But the mean is not 0 points, near 5%, the overall right shift:

After statistics, theaverage price of 2014 years is 40125/flat, 2015 is 42535/flat. The proportion of price increase 5.64%. In other words, a set of 3 million of the house, an average increase of about 160,000.

We list 100,000 yuan below the unit price, 2015 years in the community of more than 20 sets of second-hand housing prices ranked the top ten of the community:

How did the land price rise so much? Even in Beijing, 70,008 of the price is already a mansion, but the land of the house, is generally ordinary residential.

The reason is also the school District room, Haidian District teaching reform makes this side of the house changes greatly. Land in the east of the community built an experimental primary school, the primary school can be a helicopter 101 high school, 101 secondary school in the north side of the Upper West Side, is this 9-year helicopter temptation led to the region's soaring prices. Poor parents heart!

Of course, there is a price increase price: basically, the reduction of the community are in the non-central city, such as the cherry Blossom garden is in Shunyi.

Conclusion

the 5% per cent increase has already indicated that 2015 years have been warmer than the 2014 price. There are also a small number of suburban neighborhoods to reduce prices. of course, such gains are not the same as the previous rocket-like price increases. To be sure, in cities like Beijing, houses are almost impossible to plunge. But the future thing, who knows?

There are tens of thousands of scholars, Kota Kinabalu is the world's most smile solutions!

The article has been very long, so there is no more content to include. We also did the following things:

• Analyze the reasons for the price increases and visualize them on the map.

• In the same community, different house prices vary greatly, even to 20,000 yuan. Why does this happen?

• Calculate the added value of the house, based on the surrounding schools, hospitals, shopping malls and other places.

• Try to predict the future price trends in different neighborhoods.

Appendix:

Chain home in the last year there are about 7W data, this year's sale of secondary housing has reached 10W sets, but how much water in these listings? According to the 2014 data according to the number of check repeat: more than 12,000 sets of houses appeared two times, nearly 5,000 homes appeared three times, even a set of houses appeared eight times. Which water is conceivable.

At the same time, the 2014 Web page data will also provide geographical coordinates information, 2015 does not exist, so the text is related to the geographical information of the chart is 2014. In addition, although the location of the house description is very detailed, but the intermediary will not tell you this is the number of floors. Only the total floor height of the building is provided. The reason is self-evident.

Similarly, there is a problem with the accuracy of the data. A lot of houses are priced at 10,000, 20,000, and are obviously randomly marked. There are also parts of the high price outrageous, such as 880,000/flat. The data has been sifted out before processing. Avoid interfering with the analysis results.

• Data acquisition uses the crawler tools developed by the author in C #.
• Data cleaning ETL uses the tool software developed by the author.
• Data analysis uses Python.
• Matplotlib and Seaborn are used for visualization.
• The Heat force diagram uses the Baidu Map API, 0.01 degrees latitude to a sub-region, calculate the average value as the current region of the price/second-hand quantity.

Have any questions, welcome to the discussion!

Using data to speak: Data behind Beijing house price data