Good morning, the company I work at uses IBM power chipset and the coreforecast package isn't compatible with it. All of the other packages seem to work just fine and we are able to use pre 1.0 versions since they don't have the coreforecast dependency. I was going to open a github issue, but just wanted to know which repo you wanted me to open it under. Thanks for your help.

I’m new to the Slack community and excited to be here! Could anyone please share their insights or predictions for BTC price over the following timeframes? 4 hours 1 day 1 week Looking forward to your thoughts. Thanks in advance!

Hey, for a time-series data input to StatsForecast or NeuralForecast, does the 'ds' column need to be in date format or can it be consecutive numbers?

Hi there, I am trying to access Nixtla through an Azure deployment. Up until yesterday morning it was working fine but now I just keep getting timed out by the nixtla client?

Hi all, my Prediction Intervals seem very narrow when using Linear Regression in MLForecast, most probably the issue is in the user 😄 My y-value is in range 100k - 200k but the prediction intervals even with 95% level are still very narrow, example below:

lr_pipeline = make_pipeline(

LagPromotionScaleTransformer(

group_id_column="Market",

event_columns=["Mechanic 1 Detailed", "Audience 1"],

lags=[1, 2, 3, 7, 14, 21],

),

LinearRegression()

)

mlf = MLForecast(

models={"lr_pipeline":lr_pipeline},

freq="1d",

lags=[1, 2, 3, 7, 14, 21],

target_transforms=[LocalMinMaxScaler()]

)

cv_results = mlf.cross_validation(

df=features,

n_windows=10,

h=21,

static_features=[],

prediction_intervals=PredictionIntervals(n_windows=2, h=21),

level=[0.8, 0.95]

)

y	pred	lo_95	lo_80	hi_80	hi_95

153595	145229	145173	145182	145277	145285

138468	140029	139993	139999	140059	140065

143796	137528	137337	137367	137689	137719

Most often the y is outside the lo_95 and hi_95 as well as the lo_80 and hi_80. The base model performs reasonably well in terms of MAPE (~15%), but the prediction intervals do not seem to reflect the actual variance in errors. Is this expected behavior for predict_interval with linear models?

Good evening. Just an off-topic question, the slack service requires a subscription in order to view messages older than 90 days but I don't want to purchase one. Do you consider moving to or creating a community on Discord? They don't have such limitation and (I believe) it's more common for IT professionals.

@José Morales Any update on Azure issue?

What are the Python requirements for installing NixtlaClient (for TimeGPT)? I didn't find it in the online documentation

Hi, we're building something that's complementary to TimeGPT. Where's the best place to reach out about it?

hey all! can we use Nixtla for demand forecasting? think of

N

products with historical prices signal:

{p1(t=01.01.2024), p1(t=01.02.2024), ... ,p1(t=31.12.2024)},
{p2(...), ... }
 ...
{pN(t=01.01.2024), pN(t=01.02.2024), ... ,pN(t=31.12.2024)}

(sorry for the rough notation there, hope it's clear) what might be an apropriate framework to: 1. predict the future sales (say, 2025) of these N products: p1(01.01.2025) ... pN(31.12.2025) 2. predict the sales of new products pN+1, pN+2, ...

Hello we build or run anomaly detection in our company for millions of time series in one run and want to improve our ad system I had experimented with nixtla s timegpt a) it wasnt as great or too much ahead of other models it just saves effort i felt b) the biggger reason being its too expensive and doesnt make sense to pay 4 dollars per 1000 API calls when u want to run anomaly detection on a million time series (we would want something free ideally) Im exploring and would like to get an opinon if autoarima is a good model or going ahead with a combination of models Have you guys worked on this kind of stuff basically model selection and anamoly detection at scale

Hello folks, I hope this the right channel for this. I know Nixtla is a Mexico-based startup I wanted to know if folks know any good meetups, Luma events, conferences or networking sessions happening in CDMX next month. I’m going to CDMX on the 28th and I’d really like to know the AI / ML / Startup ecosystem. Cheers everyoone!

Hi folks, Is there a implementation of Pinball loss in nixtla or similar loss function for probabilistic forecasts?

Not sure whether it's just me or applies to others, but recently when I run the nf models in Colab (cpu) they run very slowly. Anybody else is experiencing this?

How to do hierarichal forecast for a grouped spark df i notice that u provide the reconicle method for one Hierarchical series But how to do i apply this in a grouped series eg one set is ['country,'brand','device_no'] another is ['country','device_no','state'] can the value hierarchy be maintained across these trees as well ie value for device no S23 must be same after rollup in tree 1 as well as tree 2

My team and I have been working with Nixtla for several months now and we feel like there must be a better solution for how we are backtesting using the cross-validation method. We currently are using a retail calendar which creates non uniform window sizes for our backtest, we hacked together a way to still use the cross-validation method but we are now running into issues with our exogenous variables. We thought being able to pass something like a list of dates for the cutoffs or something similar would work, but I wanted to reach out to the pros to get some advice!

Hi everybody. Are there helper methods to transform a general timeseries dataframe to Nixtla's unique_id/ds/y dataframe? My dataframe is quite large (20k x 100) and for every predicted step I'm updating it with the real value calling .concat and predicting the next step inside a 5000 loop iteration. Even Pandas documentation alerts regatding .append and.concat methods, how inefficient they are. The dataset manipulation is taking 99% of the time, my gpu is mostly idle. I'm stuck trying to figure out a structure to efficiently store my data and updating it to convert each iteration to the unique_id/ds/y dataframe. Any idea will be greatly appreciated.

Hello team, the Statsforecast packages has the option of direct or recursive forecasting? Is this possible?

hey guys. a question on normalizations and feature encodings: are those calculated at on raw df at the beginning, before training loop, or on a batch of data? something like: Beginning:

def fit(df,...):
    df = feature_encoding(df)
    df = normalize(df)
    for X,y in dataloader(...):
        pred = model(X) ...

in the training loop:

def fit(df,...):
    for X,y in dataloader(...):
        X = feature_encoding(X)
        X = normalize(X)
        pred = model(X) ...

Hi All! I am happy to contribute to the Nixtla Community open source projects, but I do not know how the issues are prioritized. At the moment the issues look like they are of equal importance. If I want to help how do I know what is most pressing?

Hello I'm working with a weekly freq time series of format Year-week_number, is there a way of using this data inside Nixtla? I tried to use an int freq but i get a lot of errors, and can convert directly to date because i get wrong dates.

Hi guys, i'm interested about the transfer learning for time series. Last time i watch Max's presentation it still on progress. Does it work? is there any paper and article that shows how is it performed against other dataset?

Hi guys! Regarding the

plot_series

function in utils, can anyone tell me why there's a gap between the training data and the predictions? It does not bother me much, but some people might find it weird when I'm presenting. Is there any way to get rid of that gap betwwen the series?

Hi guys, i'm trying to make a sales forecast for grocery store per sku and i'm thinking about adding a new column called past_30_days_sales as exogenous variables, what do you guys think about this? will it work in general?

Hi everyone. A very basic question I'm having just now. I'm providing datasets for neural forecast models training and predicting. My datasets all have many value columns, ie, after I convert to Nixtla's expected long format, these datasets always have many unique_ids. My question is: are the models considering them univariate or multivariate? My concern is: are the models assuming I'm expecting them to analyze any correlation between the series? Or no matter if I train each column individually or all together, the prediction of each one's next step will be the same? A final observation: according to this, every models I'm using is univariate.

Hi guys, i wonder if 36 monthly sales data points for thousands of SKUs is sufficient for training the NBEATx model, is it enough? i've tried it but it doesnt work, the prediction is bad

Hi, a general question about the implementation of the fourier transformation. I can set the freq and season length. But does this has to match the format of my timestamp or can I have daily data and set 'W' and season length 53 and it will give me valid fourier terms for weekly patterns? When I plot the data it looks to me that I first have to transform my data into the respective time aggregation and then I can run the fourier transformation only for this freq. or am i wrong here and i can have my daily data and get fourier terms also for week and month directly?

Hyndman’s new book now includes a section on conformal prediction https://otexts.com/fpppy/nbs/05-toolbox.html#conformal-prediction-for-distribution-free-prediction-intervals 🔥🔥🔥

People are welcome to feature cool use cases for Nixtla ecosystem beyond documentation for my new book https://valeman.gumroad.com/l/MasteringModernTimeSeriesForecasting dm me here or on LinkedIn and if your usecase (incl dataset) is interesting I will feature it in the book and give full credit.

Hello, thanks for supporting open source! Not sure if this is the right channel to post a subject regarding datasetsforecast but I just wanted to know why is there an "OT" time series in the end of all of the datasets comingo from longhorizon2. Is it the target prediction time series? And the rest would be the exogenous ones? Thanks in advance.