# How to build a prediction service on top of streaming (kafka) input data

**URL:** <https://discourse.numenta.org/t/how-to-build-a-prediction-service-on-top-of-streaming-kafka-input-data/3430>\
**Category:** NuPIC\
**Created:** [February 6, 2018, 4:05am UTC](https://discourse.numenta.org/t/how-to-build-a-prediction-service-on-top-of-streaming-kafka-input-data/3430 "2018-02-06T04:05:58Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![makeawish](https://avatars.discourse-cdn.com/v4/letter/m/c6cbf5/32.png) [@makeawish](https://discourse.numenta.org/u/makeawish)\
**Post date:** [February 6, 2018, 4:05am UTC](https://discourse.numenta.org/t/how-to-build-a-prediction-service-on-top-of-streaming-kafka-input-data/3430/1 "2018-02-06T04:05:58Z")

</div>

Hi.  
I have streaming input data (CPU/memory metric) and I want to build HTM on top of it and enable continuous learning from the streaming data and predict future values of the metric.

Can someone please let me know if HTM integrates with streaming services and if possible, please point out any example of such implementations?

Thank you!

---

<div class="post-metadata">

**Author:** ![rhyolight](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/rhyolight/32/3922_2.png) [@rhyolight](https://discourse.numenta.org/u/rhyolight)\
**Post date:** [February 6, 2018, 2:44pm UTC](https://discourse.numenta.org/t/how-to-build-a-prediction-service-on-top-of-streaming-kafka-input-data/3430/2 "2018-02-06T14:44:41Z")

</div>

Here’s a search that turns up some things you should read: [nupic streaming data](https://discourse.numenta.org/search?q=nupic%20streaming%20data).

---

<div class="post-metadata">

**Author:** ![psingh22](https://avatars.discourse-cdn.com/v4/letter/p/96bed5/32.png) [@psingh22](https://discourse.numenta.org/u/psingh22)\
**Post date:** [May 5, 2018, 7:12pm UTC](https://discourse.numenta.org/t/how-to-build-a-prediction-service-on-top-of-streaming-kafka-input-data/3430/3 "2018-05-05T19:12:17Z")

</div>

Hi @rhyolight, thanks for guiding me in the right direction. I have finally managed to integrate HTM and build a project on that. If anyone is facing any problems with HTM-Kafka integration,please checkout my code.

> **[gitPratikSingh/Real-time-anomaly-prediction-in-distributed-systems](https://github.com/gitPratikSingh/Real-time-anomaly-prediction-in-distributed-systems)**
>
> Using Hierarchical Temporal Memory for real time anomaly prediction in distributed systems - gitPratikSingh/Real-time-anomaly-prediction-in-distributed-systems

---

<div class="post-metadata">

**Author:** ![rhyolight](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/rhyolight/32/3922_2.png) [@rhyolight](https://discourse.numenta.org/u/rhyolight)\
**Post date:** [May 5, 2018, 9:56pm UTC](https://discourse.numenta.org/t/how-to-build-a-prediction-service-on-top-of-streaming-kafka-input-data/3430/4 "2018-05-05T21:56:24Z")

</div>

Holy cow, that’s great! Thanks for the contribution to our knowledge base!