Lets presume that collector (picker) is visually controlling a strip of 1 m width in front of him. Knowing the polygon dimensions I could easily calculate what surface was actually seen. Surprisingly it is not so few – I calculated all 373 polygons and got:

average seen surface is 13.2%, median 12.5%, min 9.45% and max 17.5% of the whole surface. By multiplication of finds number/weight I could asses amount of finds present on actual surface.

Probability of 1 piece is given by percentage of cover (e.g. around 12/13%). Consequently I could multiplicate this percentage if more sherds are on the surface…

Hm, this could not be calculated or stated per se. The only guideline would be a pottery dating – if it is of the same age (e.g. late neolithic), I could assume of the same origin (e.g. one neolithic pit). But nothing guarantees this and this is a weak point.

Well, this could be the core of my work. It means to:

a) find the polygons with the narrow-dated assemblage in several consequent seasons

b) compare the amount of samples

c) get the preliminary “visibility” coefficients

d) try to calculate what influences this visibility

e) test the result on a different polygons/different material from the same polygon

“Statistics theory falls into two camps, frequentist and Bayesian. Frequentist is older and solid. Bayesian is newer, more flexible, and more exciting.”

More about this in a nice George Casella presentation. He mentioned the “likelihoodists” as well but did not concern about them.

frequentists stats

OK, lets play first with this one. Why not descriptive? Because I was sampling and data acquired are only the portion of all data available. Why first? Cause I know nothing about sexy Bayesian… :).

multivariate

Regarding multivariate stats one has to keep in mind:

they are generally divided into **clustering** and **ordinal** types. While clustering groups objects together according similarity in variables, ordinal is arranging them in graph according most valuable (significant) variables

all multivariate methods try to reduce the number of dimensions by searching those most valuable and defining the (hidden) dependencies among them (simplification of view on data)

table of objects with variables (dimensions) is always a starting point

we could analyse relationships **among objects** or **among variables**

every method has its limitation regarding the nature of data (e.g. quantitative, qualitative etc…)

every dataset

]]>

Zdroje prezentace, tak, jak jsme si popovídali o mapách Chobolic a Liběšic v březnu 2018, jsou veřejné a online:

- archivní mapový portál ČUZK
- mapový portál Chartae Antique
- Oldmaps laboratoře geoinformatiky Ládi Brůny
- vojenský archiv leteckých snímků z Dobrušky

Byť jsou mapy dnes všudepřítomnou a standardní pomůckou, není tomu tak dlouho. Vzpoměňte si na mapové aplikace v telefonech, které tak samozřejmě používáme – kde byli před pouhými deseti lety? Vytvořit kvalitní mapu (prostorovou projekci vpírající se našim přízemním smyslům) je poměrně nákladná věc a proto to musí mít adekvátní důvod (a rozpočet). Není divu, že k takovýmto drahým řešením se uchylovala až organizace vyššího řádu, zejména stát, který byl motivovám především pohnutkami daňového censu a vojenské organisace. Ergo – první mapy byli především vojenské a berní.

Mapy jsou vždy tematické – popisují určitý úsek skutečnosti dle záměru, pro který mají sloužit. Vojenské mapy budou ergo zobrazovat pohoří (ať se tady s trénem netáhnem), řeky (ať víme, kde bude brod a kde jsou ti prokletí Saracénové) a vesnice (ať víme, kde vyplundrovat, nebo aspoň ustájit). Daňové mapy jsou mapami censu – kde co který strejda má a my mu již vysvětlíme, co jsou berně…

Pro objektivizovaný a formalizovaný náhled krajiny “sezhora” je v geografii často používá pojem Gods view (Boží náhled, neboli Boží oko). Dost výrazně kontrastuje s přirozeným lidským náhledem na krajinu (okolí), tak, jako ji vnímáme každý den. Obr. 01. a 02. tento rozdíl ozřejmují. Jak asi vnímá své okolí takový brouk žijící v trávě? A jak améba ve fyziologickém roztoku?

Přechod mezi přirozenou lidskou perspektivou (“ze strany”) a vertikální mapovou ortoprojekcí (“Boží pohled”) je tzv. “kopečková metoda”. Je typická pro nejstarší novověké mapy – prostorové dispozice zobrazuje mapově, ale vybrané typy objektů (vyvýšeniny, městá a pod.) ze šikmého (izometrického) náhledu. Tímhle začínáme i naše mapičky.

“Kopečkové” mapy se objevují vesměs od 18. stol. (pro dotčené území), jejich výtvarnost a kvalitu posuďte sami. Na všech je zobrazena dominanta Sedla (Gesetz) a jedná se o výrezy obecných map v “Leitmeritzer Kreis”. Mapy mám doma oskenované v plném rozlišení, kdyby si to někdo chtěl přijít kouknout, není problém, prezentované obrázky jsou totiž výřezy.

Vojenská mapování jsou speciálky pro vojenské účely. Dosti obšírně jsou popsány a známé, proto nebudu nosit dříví do Atén a sovy do lesa… Pro oblast Liběšic/Chobolic je k dispozici 1. a 2. vojenské mapování, ne však 3.

Stabilní katastr je monstrózní císařské dílo pozemkového censu v celé monarchii. Jeho obrovskou předností je, že:

- je poměrně přesné (dodnes podklad pro parcelaci)
- je podrobné (až do úrovně plodin na každé zahradě; zachytává každá Boží muka u silnce)
- jsou k dispozici obsáhlé údaje kolem (vlastníci, výměra parcel, měřičský operát a pod…)

Celkový náhled na Dolní Chobolice na Stabilním katastru jasně definuje rozsáhlost vesnické držby pozemků, lánový způsob parcelace (dlouhé nudle polí). Výřez na intravilán obce pak zase umožňuje stotožnit jednotlivá stavení s dnešními baráky. Přikládám i zajímavou srovnávací tabulku výměr polností Dolních Chobolic v rozmezí jednoho století. O stabilitě vesnického prostředí svědčí to, že se moc nezmněnili.

Mladý národní stát vzniknuvší po Velké válce měl příliš velké personální i fiskální problémy na to, aby se pouštěl do nějakých velkých mapových děl, srovnatelných se Stabilním katastrem. Levnější a efektivnější byla cesta letecké forogrammetrie (svislé snímkování z letadla), kterou byla republika pokryta v pravidelných časových intervalech. Nevýhodou fotogrmmetrie je absance oné mapové tematičnosti. Z Chobolic máme k dispozici snímky z r. 1938 a 1946.

Proč jsme tohle napsal?

- Protože jsme se sešli u nás doma nad mapami a chtěl jsem, aby jste ty informace měli někde pěkně po kupě, až se k tomu budete chtít vrátit
- Je jisté, že kromě uvedených “klasických” mapových děl existjí i parciální (malé) mapy, které sloužili jen ke konkrétním účelům (třeba rekonstrukce části vesnice; projektové podklady a pod.). V případě, že na takové narazíte, dejte mi vědět. Uděláme mapovou sbírku Liběšic a okolních vesniček.

]]>

Uncle Joe knew where you lived (and what is a color of your carpet) as John Davies clearly showed (part 02 is here).

The suspiciousness and paranoia of Soviet regime led to bizarre results in “public” mapping – maps were inaccurate and occasionally completely false because of military areas and objects of strategic importance (to fool western spies “present everywhere”). Quite contrary, secret maps of foreign countries were of highest quality and graphical details (see this specimen of San Diego, or this Cold War Russian maps of San Francisco). Every general know the maps are crucial for the planning of tank routes in expected epic conflict between bad capitalistic world and the “Camp of Peace”. These detailed maps were secret form 30ties of 20.th century till today and were directly under NKVD administration. Exhaustive explanation and description of their ude is available also in this article and here on a blog of Wired magazine.

OK, and now the most important info – these maps are to be downloaded here. :)

]]>Spatial interpretative division (see white lines) corresponds to our thoughts of early medieval villages/towns as we know them from Europe (even with social stratification by local nobility – see “Zamek” and lots division). Enjoy this simple naivistic ethographical attitude :)

brama = gate

rynek = square

zamek = castle/stronghold

mur = wall

Enjoy and having some additional link (sources), please be so kind and post them here, thanks.

]]>

Sharing on public although in Czech language. Laughing a lot reading this today. Download link – archaeolgy_CAD_single_context_recording

]]>`(defun c:zmen_hladinu (/ mech1 ciel1 pocet_mech1 pocitadlo var1 nazov_ciel1 var2 extrakt1 hladina)`

; local variable defined

(graphscr)

;

(prompt "\nSelect objects to be changed: ")

(setq mech1 (ssget))

; Set of object that will be changed

(setq pocet_mech1 (sslength mech1))

; we have number of entities in mech1

(prompt "\nSelect object in target layer: ")

(setq ciel1 (entsel))

; We have 1 entity stored

;

(setq nazov_ciel1 (entget (car ciel1)))

; we got the name of entity encoded in CAD db...

(setq extrakt1 (assoc 8 nazov_ciel1))

;... and extracting the name of a layer indicated by "8"

;

(setq pocitadlo 0)

; getting counter to zero

;

; starting loop for all objects in mech1, extracting and

;changing the value of layer attribute

(repeat pocet_mech1

(setq var1 (entget (ssname mech1 pocitadlo)))

(setq var2 (assoc 8 var1))

(setq hladina (subst extrakt1 var2 var1))

(entmod hladina)

(setq pocitadlo (+ pocitadlo 1))

)

(princ)

)

1. 373 sampled polygons; 353 of them walked/sampled once at least (= 20 polygons were not suitable for sampling overall), 348 of them sampled 2 times at least, 337 of them sampled 3 times at least and 222 of them sampled 4 times (in all 4 seasons were “sampleable”). These 222 polygons (59.5% of all polygons) are of mine eminent interest.

Wanna SQL query how to get this?

/*vybira jen polygony, ktere byli zkoumany ve vsech 4 sezonach*/

SELECT COUNT (DISTINCT (“id_polygon”)) FROM tab_polygony_sezony ou

WHERE (SELECT COUNT(*) FROM tab_polygony_sezony inr

WHERE inr.”id_polygon” = ou.”id_polygon”) > 3;

Polygons are of various area considering the fact I have delimited them according plots/fields boundaries as I wrote in diss 01. Fig. 01. shows the frequency histogram for a polygons areas, it is clear then most polygons had 4000-5000 square meters.

We have 11 999 finds altogether of 63 612 g weight; Fig. 02 showing the number and the weight for every season, Fig. 03 and Fig. 04 showing the types ratio of artifacts.

Wanna SQL for this?

/* vaha nalezy po typech – window funkce*/

SELECT DISTINCT(material), SUM(“vaha”) OVER (PARTITION BY material) AS Vaha_ker

FROM tab_finds;

Pottery is most the most often kind of find, there are 9542 pieces of weight 45.3 kg.

Average weight for a find is 5.3 g, the average weight of ceramic fragment is of 4.75 g. However the typical weight (median) is about 1/2 g of pottery fragment as Fig. 05 indicates (the frequency of pottery weight). The average length of a ceramic specimen is 24.5 mm, the average width of 20 mm.

The dating of surface collection samples is usually tricky because of fragmentation. However we were able to date the pieces at least generally (e.g. prehistory; recent etc.). Fig. 06 shows the dating of all finds, Fig. 07 the dating of a ceramics exclusively.

It seem both charts are in positive correlation. Well, there are a few things suspicios – the high ratio of prehistoric finds, almost absent protohistory (Roman/barbarian) and quite sparse middle ages fragments.

OK, lets load everything into GIS and get some cartograms of distribution (“where what”, concentrations and so on). I will start this with ‘naive’ attitude and just push the data as are into GIS.

First of all I wanna see the distribution of pottery in distinct seasons (season per season) assuming the distributions have to equal in some extent (I simply presume every season shows the same “sites”, more or less…) – figs. 08 – 11.

Wanna SQL for this?

/*asking for pottery weight divided by polygons and for just one season – change season number in where clause */

SELECT DISTINCT(id_polygon), datation, material, sum(“vaha”) OVER (PARTITION BY id_polygon) AS Vaha_ker

FROM tab_finds

WHERE material = ‘keramika’ AND season = 1 AND datation BETWEEN 1 AND 1999

ORDER BY id_polygon;

Although it seems we found the correlated concentration (somewhere southeastern part of trajectory), a closer lok to this section is not so clear – figs. 11 – 15.

So, better than looking on colored circles I would throw the section of data in some objective formula. From the polygon 290 until 373 I took the absolute values of prehistoric pottery weight (fig. 16) and feed the correlation algorithm. The results are following (fig. 17):

1. there is weak negative correlation among the season 1 and others

2. there is no, or weak correlation among season 2 and others

3. there is some, but not strong correlation between season 3 and 4

OK, that’s bad, mainly the negative correlation among the season 1 and others. What about, lets say, medieval sherds? They are more numerous so this primitive statistics would be more reliable. Distribution of medieval sherds on polygons 290 – 373 is on figs. 18-21.

OK, let’s throw an eye on correlation tab – figs. 22-23.

The results are pretty bad – the best is “no correlation” but the most is negative correlation among seasons. Strange – every season shows quite different distribution. Time to lose my nerves and load all available data to all polygons, based on material weight. I omit the season 1 while giving most negative correlations with the rest. While the whole map would be unclear and the trajectory is linear, I could use a simple 2-dimensional chart (X axis of polygons and Y axis of all material weight) – Fig. 24.

What about the correlation among curves? Is there patterning in all 3 seasons distribution? Hard to say (or to see) so lets order one curve ascending and reorder another curve along to see possible correspondence – fig. 25.

Well, no way. There is no common tendency and the overall distributions are not in the correlation. Starting to feel hopelessness I would jump to more sophisticated statistical attitude…

In fact the crucial questions about the aims and a future course were articulated in well known paper of Martin Kuna. I recommend to see the conclusion for clear articulation of problems.

1. We sampled in Easter Bohemia where intensive agricultural processes took place for decades (communists colektivisation before all). We would presume the huge reduction of archaeological record. How huge? Is it possible to quantify the degree of reduction?

2. We sampled the same space many times (four times) under the comparable conditions. It quickly became clear the artifacts density (and quality) differs among seasons. What are the principal attributes of surface visibility in various conditions?

3. Regarding the above questions – is there possibility “to believe” the results of surface sampling by modeling the spatial distribution of archaeological components? Under what conditions? (This could be articulated in vulgo “Was Slavomir Vencl right in his famous paper casting doubts upon the reliability of surface artifacts scattering?”)

]]>A slighly different resuts from 2014 published here. ]]>

The sampling was performed along the motorway project exclusively – along the axis at lemgth of 37 km and width of 50 m (fig. 01 on lidar shaded terrain background); in a quite monotonous landscape of eastern Bohemia (fig. 02). Generally speaking, the whole shape of sampling trajectory is not optimal while I would prefer more “quadratic shape” – to give more to width and less to length. The danger of the shape is clear from fig. 03. Presuming the hypothetical regular settlement network (blue dots) there is a good chance to avoid that and to get an empty result (this should be elaborated a little more by sampling theory with the basic question: “how reliable is the sampling of this shape”). Another negative element is the non-random nature of he trajectory (avoiding the current inhabited areas). And for the third – although the extent of 37 km seems colossal, regarding the width of 50 meters we get only 1.45 square kilometers, or the square of a side less than 1.2 km; or even worse – the blue rectangle on fig. 03 is covered only by 0.24%. And that is not reason to be enthusiastic….

The area is almost completely under the plough, the eastern Bohemia is an arable Czech granary (Figs. 04-06 show the trajectory on maps section from 1850 until now). This allowed us to hope we would continuosly collect the artefacts on the field surface.

The whole trajectory was divided into sampling polygons to a) divide the space into smaller analytical pieces and b) to respect the fields boundary (the idea is that different surface nature allows different artifacts visibility so we tried having the one polygon of the same visibility value). You could see the polygon numbering and delimitation on fig. 07. Approximate length of a polygon was 100 m and the approximate area somewhere about 4200 square meters. And now the important thing – we walked/collected this trajectory 4 times in different vegetative (agricultural) seasons: autumn 2010, spring 2011, autumn 2011, spring 2012. Of course we could not sample all 373 polygons in all seasons due to surface conditions (still growing and dense vegetative cover, not accessible etc…) but 222 that are of my eminent interest.

A few practical remarks – we normally proved to walk the whole trajectory in 5 days (= 7 km a day); all 4 sampling seasons took us 23 terrain walking/sampling days. The evidence of the artefacts done at the end of every polygon was most time consuming operation. The worst thing was returning to our cars at the end of every sampling day :)

.

There is a detailed methodology of surface sampling in Czech republic, mainly that of Martin Kuna school. I have been relying on his approach using the following instruments:

1. sampling method based on traverses/lines in each polygon

2. using the terrain forms to gain quantitative/qualitative comparable datasets

3. stable walking tempo and a stable field team

4. the collecting of all artefactual types of finds (“total sampling”)

While the width of 50 m having approximately 5 people for sampling, we kept 10 m distances and constant walking/sampling tempo. The aim was to get the reliable samples of each polygon – I am pretty sure we did not catch all artefacts on the surface but keeping above mentioned rules leads to comparable datasets. The team was stable and had the experience from previous surface sampling (fig. 10)

.

are used mostly for formalised description of polygon surface characteristics and data output consistency. It was obvious from the beginning that the visibility of the surface would have the decisive role in the quantity of finds so keeping this kind of record should be of an eminent interest.

While the finds type (pottery, stone etc…) and its dating is not important for the evaluation of the ertifact scatters, we have been collecting everything of artifactual nature.

]]>Nothing extraordinary – there are thousands of such calvaries disseminated in Czech lands, mostly near crossroads, roads and memorial places.

My neighbour Pavel has been making some earthworks (cleaning the road ditch) 30 metres aside the cavalry and has found a rectangular shaped stone sunken in a mud. After a while (and some cleaning from clay and debris) it became clear that a missing part of cavalry was found. The stone fits on the top of cavalry torso perfectly and only its color indicates the stone was left a few years in the soil and rubbish (it is sandstone and even after proper cleaning this piece remains of “muddy” grayish undertone; Fig. 01 – 02).

We spent a few hours of sniffing for information next week. Eventualy the puzzle was successfuly finished:

Sometimes in fifties (a few years after the Communist coup d’état in 1948) there was brutal general movement against Catholic Church in compliance with perverted Marx’s “religion is the opium of the people” statement. Besides the monasteries closure and churches desacralisation the country monuments were physicaly destroyed as well.

“Our” calvary was destroyed by railway employees who fulfilled their duty with little care and threw the calvary upper part in nearest ditch. This part came to the light after 60 years of lying in the mud (Fig. 03 – 05).

I did a small 3D modeling of an object still waiting for a complete reconstruction.

And why is this post in thinking stratigraphy category? It is very sparse to find the whole way of a standing structure part from its original location to archaeological deposition and vice-versa.

]]>1. While Civil surface is a spectacular (Autodesk internal) object, the export and conversion is not straightforward.

2. Explode surface 1x to achieve the block and 2x to get a bunch of 3D faces

3. Save the file as .DXF

4. Open .DXF in FBX Converter and convert .DXF to desired format (.OBJ preferred)

5. .OBJ could be opened and edited by Meshlab. Works both with BLOCK and 3DFACES got by initial exploding

1. while those Peterpuffers from Adobe left the 3D support of Acrobat gone, there are 2 main ways to get it:

a) use 3rd party solution – e.g. Haru PDF library

b) use the older Adobe Acrobat 2009 with 3D PDF support and import .stl object into

c) import AutoCAD model into Bentley View and print 3D PDF – more straighforward

This content is password protected. To view it please enter your password below:

]]>

Nothing has changed – except those tourist information tables and the beer prices in Rovinj vicinity. The expedition of me, Alsu, Max and Šaňo via Istria trip also visited the famous Iron Age hillfort at Monkodonja (it: Moncodogno; lat: Cydonia oblonga). Well, famous from my point of view because of my engagement on Ljubljana University years ago – slovenian department of archaeology led the excavation seasons from 1997 till 2007 here (OK, some Germans and Croatians along :)). The famous work of Carlo Marchesetti (I castellieri preistorici di Trieste e della regione Giulia, 1903) did not mentioned this hillfort (but did 300 others) and I recommend to read this paper for those of You interested in Istria castellieri culture.

After a rich lunch somewhere around Rovinj we took a trip to hillfort. The youngest member ignored the sacred archaeological site completely and prefered a short sleep before more interesting things – playing with the stones on the sea shore. Maybe was he right.

The fortress itself is situated on a not very distinct hill of elliptical shape. Someone could be surprised of a small dimension and the vertical insignificance in comparison to a) the expectation of “big fortress Monkodonja” and b) adjacent hills. Really, the hills around Monkodonja give the impression of “being among equals”. The viewshewd analysis provides the same result – the choice to start a settlement here was not influenced by its extraordinary altitude.

Present stone structures are the result of the previous archaeological excavation and consequent reconstruction (what about walls height? Derived from what?). A short fotoreport reveals the technical details of dry stone walls (is this the technique of “core walls” – carrying shell of big stones and small stones filling?) and somewhat spectacular entrance system of the gates.

There are rumours around the techniques of excavation to be “non stratigraphic” and “Pompei premised”. Having no idea about this I would only state that thematical interpretative results (see phase plan) seem to be of very high professional level. Contrary, the term “cult cave” used for the hole in the carstic subsoil seems to be nothing more than lack of fantasy and pragmatic attitude. What about “refrigerator” in pre-freons times?

Monkodonja today is calm and dead place with exception of ants and the snakes (notice the clay color – this country is also called “red Istria”) in contrast to Kokuletovica downhills. There are a lot of new houses there – maybe built of stones of the former Iron Age stronghold?

]]>This content is password protected. To view it please enter your password below:

]]>