Inflation may have peaked, but it is far from having normalized. A case for sticking with alternative data... and an open-mind.

The data shows euro area inflation volatility remains puzzingly elevated at this point, when compared with the US.

The lesser-known data in our datasets, but also: the latest on NIPI vs CPI and HICP and NewsBot App feature update.

Four different ways to track and analyze inflation turning points, using the quantified signals from our datasets and powerful language models.

A look at our news-based inflation diffusion index in the euro area which has still some way to go to normalize.

Language Models call for a mental shift, from trying to replicate existing data to creating new data we have never had access to before.

Introducing our new inflation metrics. Not just a new model or dataset, but a whole new way of leveraging Language Models. "

Car price dynamics have diverged in the economies in 2023. We illustrate how the News Balance can provide real-time international insights into sector-level inflation dynamics. "

What does ChatGPT change to the way we operate? And a look at Inflation Normalization progress.

Language Models trained to identify inflation relevant news continue to lead US food price inflation.

Inflation normalisation, Japan exception and Car price crash in China.

Uniquely among G7 economies, the NIPI suggests underlying inflation is not about to slow down any time soon in Japan, as core and wages pressures continue to build up.

Swedish and Polish news articles are now included in the Inflation and Wages NewsBots.

The innovation behind ChatGPT - the Transformers Language Models - are even more relevant than the specific chatbox application, because any text can be quantified now.

Today's language models are proving surprisingly reactive to the new environment. They help overcome the limitations of traditional data and models when volatility is on the rise.

Our inflation diffusion index based on news articles filtered by our NLP models has continued to track remarkably well headline inflation in the US.

Japanese news articles are now fully integrated in our suite of inflation monitoring data.

The NIPI has been signalling near-term inflation turning points, including the deceleration in US CPI Inflation at the start of the summer 2022.

We show how to combine the information from two of our timely inflation indicators to detect near-term turning points.

In the short term, inflation is a collection of largely unrelated micro shocks. Natural Language Processing can help process the huge amount of information required.

Energy prices dynamics can now substantially diverge from one region to the other. AMS NIPI Energy prices helps keep track of retil energy price developments in real time.

China is far from immune to the global inflation shock

We have compiled hundreds of thousands of news articles in Chinese to build the first ever NIPI for China.

To monitor inflation momentum is tricky when seasonal adjustment is an issue. Enter the NIPIs.

Timely catch of recent US CPI inflation turning points

Our NIPI has down a fine job tracking US food prices when compared to traditional data

Our new euro-area NIPI shows a marked increase in inflation pressures, within just a few weeks

The overriding consensus remains that the inflation shock is transitory: is it justified or is it a failure to conceive something else than "back to previous state" post-COVID?

Our News Inflation Pressure Indices now clearly signal a broadening of the shock, to Europe in particular

Our news tracking models have signaled a significant acceleration in US wages, since the start of 2021

Easy Python scripts to access our data

A quick look back at how our NIPI signaled US CPI turning points in the last 18 months.

Food prices jumped at the end of last year. Check what our News Inflation pressure Index had to say back then.

Comparing the News Inflation Pressure Index and Breakeven Inflation rates

Our NIPI is at a 3y high (21-Dec-2020)

Too often, a "sentiment" indicator is there for lack of a better word and, sometimes, of a proper thinking about the concept and goal.

That's the number of news processed by our algos to compute the News Inflation Pressure Indices (NIPIs) and News Volume Indices (NVIs). Check this post to know how the data is compiled and aggregated.

We open up our data.

Our news-based indicator does track US inflation which has two interesting implications 1) the information content in our NewsBot is good and 2) we can build a whole new category of indicators aimed at nowcasting economic variables using news.

Our approach to using language models in macro: supervised learning and specialised models.

High frequency analytics to track food prices.

We show how Natural Language Processing (NLP) can deal with rather complex tasks when given the right inputs.

Machine learning can considerably enhance information gathering, within pretty much any investment process.

The most important step in any Machine Learning project is to define the use case. In macro finance, inflation ticks all the boxes.