diff --git a/sigs/dx/README.md b/sigs/dx/README.md
index b3455c7..001e56b 100644
--- a/sigs/dx/README.md
+++ b/sigs/dx/README.md
@@ -42,3 +42,4 @@ This is the working repo for the Developer eXperience Special Interest Group (SI
 - [Tuesday, October 12, 2021](./meetings/010-20211012.md)
 - [Tuesday, November 9, 2021](./meetings/011-20211109.md)
 - [Tuesday, November 23, 2021](./meetings/012-20211123.md)
+- [Tuesday, December 7, 2021](./meetings/013-20211207.md)
diff --git a/sigs/dx/meetings/013-20211207.md b/sigs/dx/meetings/013-20211207.md
new file mode 100644
index 0000000..0b7bc38
--- /dev/null
+++ b/sigs/dx/meetings/013-20211207.md
@@ -0,0 +1,24 @@
+# Tuesday, December 7, 2021 at 19:00 GMT+8
+
+## Agenda
+
+* 1 - Introduce of Developer eXperience SIG - Qi Guoqiang / Clement Li
+* 2 - Community issue developer status measurement progress - Zhao Boxuan
+* 3 - Ranking of community contributors based on entropy method - Xia Wenzong / Qi Guoqiang
+* 4 - Use keyword matching to recommend issue's comp tag progress - Cao Jie
+
+## Conference links
+
+* Meeting link: [https://meeting.tencent.com/dm/PMdIcpyW1TtC]( https://meeting.tencent.com/dm/PMdIcpyW1TtC)
+
+* Meeting record link: [Wait for update](https://www.bilibili.com/video/BV1rK4y1T7yg/)
+
+## Attendees
+
+* Shanghai Jiaotong University OpenSource Team, Yijun Sun, Naipan Zhu, Clement Li, Guoqiang Qi, Yuhang Wang
+
+## Notes
+
+* 1 - The entropy method ranking has a positive effect on balancing the distribution of different activities in the community, but when the community develops to a certain extent, it is not very reasonable to define the value of activities by the number of activities. The value of community activities is difficult to determine the fixed weight artificially, and it is considered later. Iterate on the method to make it more convincing.
+
+* 2 - Using keyword matching to set the comp tag has the problem of too many keywords, and the keywords cannot be exhaustive. There is overlap between keywords between different tags. Later, you can consider cross-checking different tags and keywords, and the same keyword is on different tags. The recommendation occupies different weights, and the label can be recommended when a fixed threshold is reached.