I have a dataframe(‘datamining2’) like this.

'{"schema":{"fields":[{"name":"index","type":"integer"},{"name":"docid","type":"integer"},{"name":"title","type":"string"},{"name":"text","type":"string"},{"name":"lineid","type":"integer"},{"name":"datum","type":"number"},{"name":"mining","type":"number"}],"primaryKey":["index"],"pandas_version":"1.4.0"},"data":[{"index":8,"docid":3,"title":"Data mining","text":" datum mining   be the process of discover pattern in large datum set involve method at the intersection of machine learning , statistic , and database system . datum mining be an interdisciplinary subfield of computer science and statistic with an overall goal to extract information ( with intelligent method ) from a data set and transform the information into a comprehensible structure for further use . datum mining be the analysis step of the \" knowledge discovery in database \" process , or KDD . aside from the raw analysis step , it also involve database and data management aspect , datum pre - processing , model and inference consideration , interestingness metric , complexity consideration , post - processing of discover structure , visualization , and online updating . the difference between data analysis and data mining be that datum analysis be use to test model and hypothesis on the dataset , e.g. , analyze the effectiveness of a marketing campaign , regardless of the amount of datum ; in contrast , data mining use machine - learning and statistical model to uncover clandestine or hidden pattern in a large volume of datum . Olson , D. L. ( 2007 ) . datum mining in business service . Service Business , 1(3 ) , 181 - 193 . doi:10.1007 \/ s11628 - 006 - 0014 - 7 ","lineid":0,"datum":0.6428571429,"mining":0.6},{"index":9,"docid":3,"title":"Data mining","text":" the term \" datum mining \" be in fact a misnomer , because the goal be the extraction of pattern and knowledge from large amount of datum , not the extraction ( mining ) of datum itself . it also be a buzzwordOKAIRP 2005 Fall Conference , Arizona State University   and be frequently apply to any form of large - scale datum or information processing ( collection , extraction , warehousing , analysis , and statistic ) as well as any application of computer decision support system , include artificial intelligence ( e.g. , machine learning ) and business intelligence . the book Data mining : practical machine learning tool and technique with Java ( which cover mostly machine learning material ) be originally to be name just practical machine learning , and the term data mining be only add for marketing reason . often the more general term ( large scale ) datum analysis and analytic \u2013 or , when refer to actual method , artificial intelligence and machine learning \u2013 be more appropriate . ","lineid":1,"datum":0.3571428571,"mining":0.4},{"index":10,"docid":3,"title":"Data mining","text":" the actual data mining task be the semi - automatic or automatic analysis of large quantity of datum to extract previously unknown , interesting pattern such as group of datum record ( cluster analysis ) , unusual record ( anomaly detection ) , and dependency ( association rule mining , sequential pattern mining ) . this usually involve use database technique such as spatial index . these pattern can then be see as a kind of summary of the input datum , and may be use in further analysis or , for example , in machine learning and predictive analytic . for example , the data mining step might identify multiple group in the datum , which can then be use to obtain more accurate prediction result by a decision support system . neither the datum collection , datum preparation , nor result interpretation and reporting be part of the data mining step , but do belong to the overall KDD process as additional step . ","lineid":2,"datum":0.4285714286,"mining":0.5},{"index":11,"docid":3,"title":"Data mining","text":" the related term datum dredging , datum fishing , and datum snoop refer to the use of data mining method to sample part of a large population datum set that be ( or may be ) too small for reliable statistical inference to be make about the validity of any pattern discover . these method can , however , be use in create new hypothesis to test against the large datum population . ","lineid":3,"datum":0.3571428571,"mining":0.1},{"index":25,"docid":10,"title":"Pattern recognition","text":" pattern recognition be the automate recognition of pattern and regularity in datum . pattern recognition be closely relate to artificial intelligence and machine learning , \n  together with application such as data mining and knowledge discovery in database ( KDD ) , and be often use interchangeably with these term . however , these be distinguish : machine learning be one approach to pattern recognition , while other approach include hand - craft ( not learn ) rule or heuristic ; and pattern recognition be one approach to artificial intelligence , while other approach include symbolic artificial intelligence . a modern definition of pattern recognition be : ","lineid":0,"datum":0.0714285714,"mining":0.1},{"index":26,"docid":10,"title":"Pattern recognition","text":" this article focus on machine learning approach to pattern recognition . pattern recognition system be in many case train from label \" training \" datum ( supervised learning ) , but when no label datum be available other algorithm can be use to discover previously unknown pattern ( unsupervised learning ) . machine learning be the common term for supervised learning method and originate from artificial intelligence , whereas KDD and datum mining have a large focus on unsupervised method and strong connection to business use . pattern recognition have its origin in engineering , and the term be popular in the context of computer vision : a lead computer vision conference be name Conference on Computer Vision and Pattern Recognition . in pattern recognition , there may be a high interest to formalize , explain and visualize the pattern , while machine learning traditionally focus on maximize the recognition rate . yet , all of these domain have evolve substantially from their root in artificial intelligence , engineering and statistic , and they 've become increasingly similar by integrate development and idea from each other . ","lineid":1,"datum":0.2142857143,"mining":0.1},{"index":33,"docid":14,"title":"Machine learning","text":" machine learning ( ML ) be the scientific study of algorithm and statistical model that computer system use in order to perform a specific task effectively without use explicit instruction , rely on pattern and inference instead . it be see as a subset of artificial intelligence . machine learning algorithm build a mathematical model base on sample datum , know as \" training datum \" , in order to make prediction or decision without be explicitly program to perform the task . machine learning algorithm be use in a wide variety of application , such as email filtering , and computer vision , where it be infeasible to develop an algorithm of specific instruction for perform the task . machine learning be closely relate to computational statistic , which focus on make prediction use computer . the study of mathematical optimization deliver method , theory and application domain to the field of machine learning . datum mining be a field of study within machine learning , and focus on exploratory datum analysis through unsupervised learning . in its application across business problem , machine learning be also refer to as predictive analytic . ","lineid":0,"datum":0.2857142857,"mining":0.1},{"index":35,"docid":15,"title":"Formal concept analysis","text":" formal concept analysis find practical application in field include datum mining , text mining , machine learning , knowledge management , semantic web , software development , chemistry and biology . ","lineid":1,"datum":0.0714285714,"mining":0.2},{"index":42,"docid":17,"title":"R (programming language)","text":" r be a programming language and free software environment for statistical computing and graphic support by the R Foundation for Statistical Computing . the r language be widely use among statistician and datum miner for develop statistical software and datum analysis . poll , data mining survey , and study of scholarly literature database show substantial increase in popularity ;   r rank 22nd in the TIOBE index , a measure of popularity of programming language . ","lineid":0,"datum":0.1428571429,"mining":0.1},{"index":61,"docid":28,"title":"Cluster analysis","text":" cluster analysis or clustering be the task of group a set of object in such a way that object in the same group ( call a cluster ) be more similar ( in some sense ) to each other than to those in other group ( cluster ) . it be a main task of exploratory datum mining , and a common technique for statistical datum analysis , use in many field , include machine learning , pattern recognition , image analysis , information retrieval , bioinformatic , data compression , and computer graphic . ","lineid":0,"datum":0.1428571429,"mining":0.1},{"index":66,"docid":29,"title":"Curse of dimensionality","text":" cursed phenomenon occur in domain such as numerical analysis , sample , combinatoric , machine learning , data mining and database . the common theme of these problem be that when the dimensionality increase , the volume of the space increase so fast that the available datum become sparse . this sparsity be problematic for any method that require statistical significance . in order to obtain a statistically sound and reliable result , the amount of datum need to support the result often grow exponentially with the dimensionality . also , organize and search datum often rely on detect area where object form group with similar property ; in high dimensional datum , however , all object appear to be sparse and dissimilar in many way , which prevent common datum organization strategy from be efficient . ","lineid":1,"datum":0.3571428571,"mining":0.1},{"index":68,"docid":31,"title":"Web mining","text":" web mining be the application of data mining technique to discover pattern from the World Wide web . as the name propose , this be information gather by mine the web . it make utilization of automate apparatus to reveal and extricate datum from server and web2 report , and it permit organization to get to both organized and unstructured information from browser activity , server log , website and link structure , page content and different source . ","lineid":0,"datum":0.0714285714,"mining":0.2},{"index":78,"docid":34,"title":"Relational data mining","text":" relational data mining be the data mining technique for relationaldatabase . Dzeroski , Saso , Lavra\u010d , Nada ( Eds . ) , Relational Data Mining , Springer 2001   unlike traditional data mining algorithm , which look for \n pattern in a single table ( propositional pattern ) , \n relational datum mining algorithm look for pattern among multiple table \n ( relational pattern ) . for most type of propositional \n pattern , there be correspond relational pattern . for example , \n there be relational classification rule ( relational classification ) , relational regression tree , and relational association rule . ","lineid":0,"datum":0.0714285714,"mining":0.4},{"index":79,"docid":34,"title":"Relational data mining","text":" there be several approach to relational datum mining : \n  Inductive Logic Programming ( ILP ) \n  Statistical Relational Learning ( SRL ) \n  Graph Mining \n  Propositionalization \n  multi - view learn ","lineid":1,"datum":0.0714285714,"mining":0.1},{"index":84,"docid":37,"title":"ROOT","text":" ROOT be an object - orient program and library develop by CERN . it be originally design for particle physics datum analysis and contain several feature specific to this field , but it be also use in other application such as astronomy and datum mining .   the late release be 6.18.00 , as of 2019 - 06 - 25 . ","lineid":0,"datum":0.1428571429,"mining":0.1},{"index":100,"docid":46,"title":"Data dredging","text":" conventional test of statistical significance be base on the probability that a particular result would arise if chance alone be at work , and necessarily accept some risk of mistaken conclusion of a certain type ( mistaken rejection of the null hypothesis ) .   this level of risk be call the significance . when large number of test be perform , some produce false result of this type , hence 5 % of randomly choose hypothesis turn out to be significant at the 5 % level , 1 % turn out to be significant at the 1 % significance level , and so on , by chance alone . when enough hypothesis be test , it be virtually certain that some will be statistically significant but misleading , since almost every datum set with any degree of randomness be likely to contain ( for example ) some spurious correlation . if they be not cautious , researcher use data mining technique can be easily mislead by these result . ","lineid":2,"datum":0.0714285714,"mining":0.1},{"index":137,"docid":57,"title":"Data stream mining","text":" Data Stream Mining be the process of extract knowledge structure from continuous , rapid datum record . a data stream be an ordered sequence of instance that in many application of datum stream mining can be read only once or a small number of time use limited computing and storage capability . \n in many datum stream mining application , the goal be to predict the class or value of new instance in the data stream give some knowledge about the class membership or value of previous instance in the data stream . \n machine learning technique can be use to learn this prediction task from label example in an automate fashion . \n often , concept from the field of incremental learning be apply to cope with structural change , on - line learning and real - time demand . \n in many application , especially operate within non - stationary environment , the distribution underlie the instance or the rule underlie their labeling may change over time , i.e. the goal of the prediction , the class to be predict or the target value to be predict , may change over time . this problem be refer to as concept drift . ","lineid":0,"datum":0.2142857143,"mining":0.2},{"index":138,"docid":57,"title":"Data stream mining","text":" example of data stream include computer network traffic , phone conversation , ATM transaction , web search , and sensor datum . \n Data stream mining can be consider a subfield of data mining , machine learning , and knowledge discovery . ","lineid":1,"datum":0.0714285714,"mining":0.2},{"index":149,"docid":60,"title":"JMP (statistical software)","text":" the software can be purchase in any of five configuration : JMP , JMP Pro , JMP Clinical , JMP Genomics and the JMP Graph Builder App for the iPad . JMP can be automate with its proprietary scripting language , JSL . the software be focus on exploratory visual analytic , where user investigate and explore datum . these exploration can also be verify by hypothesis testing , data mining , or other analytic method . in addition , discovery make through graphical exploration can lead to a design experiment that can be both design and analyze with JMP . ","lineid":1,"datum":0.0714285714,"mining":0.1},{"index":157,"docid":65,"title":"Orange (software)","text":" Orange be an open - source datum visualization , machine learning and data mining toolkit . it feature a visual programming front - end for explorative datum analysis and interactive datum visualization , and can also be use as a Python library . ","lineid":0,"datum":0.2142857143,"mining":0.1}]}'

So, I want to generate a facet altair chart (TileBars) with each row being an individual document (represented by ‘docid’). This is my current output. I want to fix the overlapping label to the left of y-axis and make it horizontal.

The code I used in generating this output is quoted below.

    charts = alt.Chart(dataminingdf2).mark_rect(stroke='black', strokeWidth=0.1).encode(
        x=alt.X('lineid:O', axis=alt.Axis(labelAngle=0)),
        y=alt.Y('variable:N', title=''),
        color=alt.Color('value:Q',scale=alt.Scale(
            range=['white','black'])),
        tooltip=[
        alt.Tooltip('variable:N', title='Column'),
        alt.Tooltip('value:Q', title='Value')
    ]
    ).facet(
        row=alt.Row(
            'title:N', 
            sort=dataminingdf2['title'].to_list()),
#             header=alt.Header(labelAngle=45),
            spacing=10
    ).transform_fold(
        query_terms,
        as_=['variable', 'value']
    )

    return charts

I referred to other posts and wanted to add header=alt.Header(labelAngle=0) but it throws an error.

stackoverflow1

stackoverflow2

.facet(
    row=alt.Row(
        'a:N',
        title=None,
        header=alt.Header(labelAngle=0, labelAlign='right')
    )

Note that the last 2 columns (or n-columns) could be different. It represents the search query terms in that visualisation and it could be ‘abc’ ‘def’ next time. It cannot be hardcoded.

Any suggestions? or any codes that can make the same output in a more efficient way?
Thanks a lot.